Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

We tend to notice infrastructure only when it fails, when a bridge closes, a train derails, schools strike, and mental health works the same way. It carries how we work, learn and care for each other, and it absorbs the shocks along the way, yet it rarely shows up in a budget until something breaks. In this edition, we look at what changes when you treat mental health and wellbeing as infrastructure.

It has also been a busy few weeks on the road: Amsterdam during the GIIN Impact Forum, Vienna last week at invest.austria, and Brussels next week. And on November 18, our Mind, Money & Impact, investors’ lunchtime series with Prospira Global returns. Plus a few other stories we think are worth your time.

📊 This Week’s Story

The Infrastructure You Only Notice When It Fails

Bridges get inspected because someone who pays for the upkeep also collects the return, through tolls or taxes. Mental health has many beneficiaries (employers, governments, families, and each of us), but no one who collects enough of the return to fund the upkeep alone. So it goes unmaintained until something fails.

What poor mental health costs

This year’s OECD report, The Economic Case for Preventing Mental Ill Health, estimates that poor mental health will reduce GDP by an average of 1.7% a year between 2025 and 2050, through lower workforce participation, absenteeism and presenteeism. Care is far from meeting need: the same report estimates that reaching universal access in the EU would take a 41% increase in mental health spending.

Why prevention rarely scales

It also finds that prevention programs in primary care, schools and workplaces are highly cost-effective, often returning more in productivity and lower health spending than they cost to run, yet few operate at scale. The reason is accounting.

When a school funds a mental health program, most of the savings land in other budgets, often a decade later: the health service treats fewer adults, a future employer loses fewer sick days, and the benefits system pays out less. Economists call this the wrong-pocket problem: the one who pays is rarely the one who saves, so nobody budgets for prevention.

The rise in need goes beyond more diagnosis

NHS England’s national survey found that the share of 8- to 16-year-olds with a probable mental disorder rose from about one in eight in 2017 to one in five in 2023. Yet England’s Growing Up in the 2020s study of nearly 8,000 pupils found that 25% met screening criteria for a probable mental-health disorder, while 94% said they felt neutral or happy with life. Researchers such as Lucy Foulkes and Jack Andrews argue that part of the apparent rise reflects greater awareness and the over-interpretation of normal distress.

In October, the UK’s independent Fonagy review looked at mental health conditions, ADHD and autism and concluded that rising need can’t be explained by changes in diagnosis alone. It recommends that services respond to need rather than wait for a diagnostic label, with more weight on prevention and earlier support.

Brain health: savings that arrive decades later

According to the 2024 Lancet Commission on dementia, roughly 45% of dementia cases could be prevented or delayed by addressing 14 modifiable risk factors.

Five Lives (France/UK), a Masawa portfolio company since 2024, builds lifestyle-based brain health programmes around five pillars: nutrition, sleep, physical activity, mood and social connection, and mental stimulation. Since January 2026, Five Lives Care has combined brain training with adapted physical activity for people with mild cognitive impairment. Helping someone stay independent for longer saves money for care systems and families years or decades from now, while today the individual pays most of the cost.

Workplaces: the employer pays and saves

At work, the gap between payer and saver is much smaller. In the Netherlands, employers must keep paying at least 70% of an employee’s wages for up to two years of sick leave, so the employer carries the cost of absence directly, and a large occupational-health industry has grown up around that obligation.

Colbe (formerly the Zorg van de Zaak network) grew out of that industry: a European network of more than 20 health and wellbeing companies with 2,500+ professionals, covering around 1.3 million employees. It describes its own direction as moving from absence management to work wellbeing, and this week it reported that mental-health-related sickness absence is up 51% since 2018, with occupational physicians recording mental health complaints in almost half of all sickness absence reports.

Quan (Masawa’s first investment, exited to TrueTribe, part of Colbe) measures more than 200 wellbeing predictors across work and personal life, flags teams at risk before problems turn into absence, and connects managers with support. Its clients have included ASML, Bynder, Informa and Fitch Group. TrueTribe began as Zorg van de Zaak’s corporate venture arm and now sits inside Colbe, so Quan’s early-warning tools now reach employers who already pay for absence.

What this means for investors

Mental health infrastructure gets built where someone already pays for the failure. Employers pay for absence today, and that’s why the workplace market is furthest along. Care systems and insurers carry the cost of lost independence, and companies are starting to build there too. The key opportunity is embedding solutions directly into existing touchpoints, like a medical consultation, a formal diagnosis, or a sick-leave filing, and partnering with the stakeholders who ultimately realize the cost savings.

Our Mind, Money & Impact session on November 18 with Prospira Global takes this up directly.

If you are building or backing companies that work where the cost of poor mental health already lands, with employers today and with care systems and insurers next, we’d like to hear from you.

And if you are building in Belgium, specifically, the Helios Foundation has just opened a call for projects on wellbeing at work, focused on what happens before absence starts. Projects need a clear link to Belgium and must cover at least two of three areas: healthy work organization and prevention, management behavior and working relationships, and knowledge about staff wellbeing. Selected projects can receive funding from €50,000 a year, and applications close on November 30. Read the call

📍 Masawa Spotlight

☕ Amsterdam | October 26–29

We had a great week in Amsterdam during the GIIN Impact Forum week, with the main takeaway that mental health is truly invisible infrastructure. At the IFI Training Day, we focused on innovative fund structures (like Masawa) and financial instruments. On Wednesday, we co-hosted a breakfast with BonVenture on whether AI can make impact investing hot(ter). In true Masawa Salon fashion, we had people moving around in deep discussion on the merits and perils of AI for impact investing. Later that day, at Philips Foundation’s Healthcare Connect, board chair Marnix van Ginneken said health needs capital, innovation and evidence at once. In mental health, the innovation and evidence are already there. Capital is what’s missing.

🎻 Vienna | November 3–5

Last week Masawa was in Vienna, including the invest.austria conference on Wednesday. Austrian startups raised €472m in 97 rounds in the first half of 2026, almost double the €253m raised in all of 2025. But the largest rounds still depend heavily on investors from abroad, which is the gap the planned Rot-Weiss-Rot fund of funds is meant to help close. On Wednesday evening, we co-hosted a cozy dinner on the LP/GP generational shift, with Bridges, Allocator One and Capnamic Ventures. What did we cover?

• Limited partners look at strategy, then ask for evidence. Participants agreed that a clear thesis gets a first-time fund into the room. LPs then test it against the pipeline, the team’s sector experience, and terms that fit the fund size.
• Family offices that invest directly still commit to funds, for coverage, diversification, and de-risking. Several told us they back funds to reach a sector or geography they cannot cover themselves. Downside risk is also part of the impact conversation. Investors asked how our Nurture Capital model reduces founder risk, not only how it creates impact.
• A generational handover is opening room for new managers. Many emerging GPs will graduate and baby boomers will pass on more of their wealth. Participants expect this to create space for emerging managers, providing more real estate and strategic capital moves into venture. In Austria, much of that wealth still sits in real estate.

🇧🇪 Brussels | November 9–10

This week we are in Brussels. On Monday evening, join us at Norrsken House for “Profit for Good”, an evening with Peter Singer and Stefan Krook on whether business can do good with its profits while still building ambitious companies. RSVP here.

If you would like to meet while we are in town, reply to this email.

🇬🇧 London | November 16–19

Next week we are in London. On Wednesday, November 18, during the Jefferies Global Healthcare Conference, we are co-hosting an invite-only LP breakfast with Zinc. Rich Andrews (Univa Health), Nick Sireau (Serenatis Bio) and Ash Ruddy (Neuly) will talk about what it takes to reach patients and paying customers. Places are limited and open to investors, family offices, foundations and LPs. Request to join

🎥 Mind, Money & Impact #2: Mental Health as Infrastructure

Mental health is shaped by the conditions in which people are born, grow, learn, work and live — from education and economic security to social connection, gender equity and a changing climate.

In the second edition of our lunchtime series with Prospira Global, Sabine will moderate a fireside conversation with Elisha London, Founder and CEO of Prospira Global, drawing on Prospira’s new report, Mental Health Integration for Impact. We will explore:

• Why mental health is essential infrastructure for impact, a cross-cutting determinant of outcomes rather than simply a healthcare category.
• How women’s health, financial wellness, child and youth development, human agency and climate resilience intersect with mental health.
• Where investable opportunities are emerging beyond traditional mental-health services, and what credible integration looks like in practice.

Wednesday, November 18 | 12–1 p.m. London | 1–2 p.m. CET | 7–8 a.m. ET | Online. Sign up on Luma for a spot.

📚 What Else We’re Reading…

🛡️ A call for founders to build AI safety institutions

Coefficient Giving (formerly Open Philanthropy) runs Project Tailwind, a call for founders to start new organizations on AI safety problems that no one currently owns. Areas include independent testing of AI systems, evidence generation and AI security, and founders can also pitch their own idea. The call is still open. Read the call

🇺🇸 Young adults’ distress is easing, but still far above 2012

A new CDC report finds that 21.3% of Americans aged 18 to 29 reported frequent mental distress in 2024, defined as 14 or more mentally unhealthy days in the past month. The rate fell in both 2023 and 2024, but remains well above the 12.7% recorded in 2012. The analysis covers more than 424,000 respondents. Read the report

📱 The FDA clears an app for the negative symptoms of schizophrenia

Click Therapeutics’ Motivista, a prescription smartphone therapy used alongside clinician-managed care, received FDA clearance in October. In a Phase III trial of 457 adults, patients improved 6.8 points on the CAINS-MAP scale at 16 weeks, against 4.2 for a digital control app. Click says it is the first FDA-authorized treatment for negative symptoms, which medication rarely helps. Read more

💸 Why we spend when we are anxious

In Madame Figaro, psychoanalyst Nicole Prieur argues that impulsive spending is often a psychological defense against anxiety rather than a lack of discipline. It ties into the financial wellness thread of our November 18 session. Read the interview (en français)

😴 Treating insomnia eases suicidal thoughts

A Nature Mental Health review of 19 randomized trials with 4,974 adults found that cognitive behavioral therapy for insomnia (CBT-I) cut the severity of suicidal thoughts by 23% right after treatment, compared with control groups. The effect at later follow-ups was unclear, so the authors see it as an addition to suicide prevention, not a replacement. Read the study

🇮🇸 School phone bans didn’t improve teen mental health in Iceland

A JAMA Network Open study used about 24,000 survey responses from Icelandic 13- to 15-year-olds between 2016 and 2023. Teens at schools that banned phones showed no consistent improvement in depression symptoms, wellbeing or bullying. Schools chose whether to ban phones, so the study cannot prove cause and effect. Read the study

🧬 Longevity goes mainstream, but the data rarely changes behavior

The a16z’s State of Longevity argues that health has become a consumer category, pushed along by COVID and by AI tools that make medical information easier to understand. In their survey of 500 US consumers, 54% use a wearable, yet 55% of those say their data has never led to a behavior change or a doctor’s visit, and 39% would use a non-invasive brain-sensing device for sleep, focus or mood. Their bet is that value moves to the companies that help people act on what they learn. Read the article

🎯 In Closing

Would you like to bring these conversations to your own city or community?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better: support the Masawa mission or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

🌉 Masawa Minute 162 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

Founders are trained to manage cash, teams and investors. Few are trained to manage the stress response that shapes every decision they take under pressure. In this edition, we look at what a world-record freediver can teach founders about that response, what the evidence says about the simplest mental tools, and why investors have a direct interest in whether founders use them.

Sabine also shares what she took from Norrsken Impact/Week in Barcelona, and next week we start a seven-week European roadshow beginning in Amsterdam with an investor breakfast co-hosted with BonVenture during the Global Impact Investing Network (GIIN) Impact Forum. Plus six other stories we think are worth your time.

🫁 This Week’s Story

Hold Your Breath: Founders’ Mental Toolkit and Why Investors Have a Stake

Christian Redl started his career as an investment banker. Alongside it, he took up freediving, and he now holds 11 world records in the sport, several of them set under ice. In February 2024 he set an Austrian record by swimming 75 metres under ice in a pair of swimming trunks. This week in Munich, where Sabine met him, he spoke on a theme that translates well beyond the water: how we think decides what becomes possible.

Freediving rests on a piece of physiology that every founder lives with.

During a breath-hold, the urge to breathe is triggered mainly by rising carbon dioxide, not by a lack of oxygen. The alarm sounds long before the body is in danger. Freedivers train to recognise that signal, stay calm while it rises, and keep making clear decisions. The ones who panic burn oxygen faster and end the dive early.

Founders face a version of the same problem every week. A missed target, a difficult board meeting or a stalled fundraise sets off the same stress response, whether or not the company is in danger. How a founder handles that signal shapes hiring, spending and strategy. We think founder mental fitness is a performance and governance variable, the tools to build it are cheap and teachable, and investors should treat it as part of how they create value.

1. The numbers founders rarely discuss

The evidence on founder mental health is consistent. In a study of 242 entrepreneurs, Michael Freeman and colleagues at the University of California, San Francisco found that 49% reported one or more lifetime mental health conditions, including depression, attention deficit hyperactivity disorder (ADHD), bipolar conditions and substance use. Including family history, 72% were affected directly or indirectly.

The pressure has not eased in Europe. Sifted’s 2025 survey of European founders found that more than half had experienced burnout in the previous year.

The cost shows up in company outcomes. Noam Wasserman of Harvard Business School, in The Founder’s Dilemmas, attributes 65% of failures among high-potential startups to conflict within the founding team. Conflict between co-founders is rarely only about strategy. It is often about how two people under sustained pressure regulate themselves and each other.

2. What the evidence says about the simplest tools

The tools with the strongest evidence are also the least expensive.

Breathing. In a randomised trial at Stanford published in Cell Reports Medicine in 2023, 108 participants practised one of four techniques for five minutes a day over 28 days: cyclic sighing (two inhales through the nose followed by a long exhale), box breathing, cyclic hyperventilation with breath retention, or mindfulness meditation. All four groups reported less anxiety and less negative mood. Cyclic sighing was the only technique that improved positive mood significantly more than mindfulness meditation, and it produced the largest drop in resting breathing rate. The trial was small, remote and ran for four weeks, so it shows what a short daily practice can do, not what it does over years.

Meditation. A 2014 meta-analysis in JAMA Internal Medicine by Madhav Goyal and colleagues reviewed 47 randomised trials with 3,515 participants. Mindfulness meditation programmes showed moderate evidence of reducing anxiety and depression at eight weeks, with effects comparable to those reported for antidepressants in primary care populations.

Structured listening. Peer formats with clear rules, in which one person speaks and others listen without advice or interruption, address the isolation many founders describe. They cost nothing and require only time and a facilitator.

Redl teaches a breathing technique that, in his words, can be recalled and used at any time in everyday life. None of these tools depends on an app, a subscription or a device. Each builds a capacity the person keeps.

3. What the toolkit looks like in practice

This is the thinking behind the Masawa Salon. Each Salon draws on a menu of activities designed for founders, investors and business leaders, sequenced from opening to closing:

• A Founder Mental Health Spectrogram, in which participants stand along a line to estimate how common mental health challenges are among founders, then compare their guesses with the data.
• A Case Clinic from the Presencing Institute’s Theory U, in which one participant presents a leadership challenge and peers help surface new perspectives through listening and reflection rather than advice.
• A Failure Timeline, in which participants chart the setbacks in their career so far and look at what each one taught them. Seen side by side, failures read as a sequence of lessons, which makes it easier to meet the next one with less self-criticism.
• An Inner Boardroom, a guided visualisation in which participants identify the internal voices that drive their decisions, such as the cautious one, the ambitious one or the people-pleaser, and practise noticing which of them is speaking when a hard call comes up.
• A Power Mapping, in which participants map their sources of power, such as capital, expertise, network or reputation, and where they spend it. For founders, this often shows how much influence they give away to investors or to a single customer, and where they hold more than they assumed.
• A River of Life, in which each participant draws their life or leadership journey as a river, with its key events, rapids, turning points and the people who supported them along the way. It puts the current pressure in the context of a longer path.

The sequence matters. Participants first see that their experience is common, then learn a skill they can use the next morning, then practise the kind of conversation most founders rarely have with their investors.

For founders who want to keep the practice going between Salons, tools exist that are built for the working day. If you do use an app, look for one that builds a skill you keep, like Regulate. Founded in Germany by Peter van Woerkum and Paul Laechelin, Regulate is a breathwork app designed for professionals, with more than 40 sessions of 1 to 60 minutes for specific moments: before a high-stakes meeting, after a tense discussion, or between two contexts. Peter led the breathwork session at our impact investor breakfast with ImpactVC in Berlin in May 2025 and at our Christmas get-together in Munich in December 2024, and participants left with trial access to keep practising.

4. The market is building mental fitness tools, with different incentives

Mental fitness is attracting serious capital. On 6 October, Flagship Pioneering launched iris labs with a $20M commitment. Its first product, iris chat, is an AI “thought partner” delivered by text message and framed around “psychological fitness”. Sword Health, which completed its acquisition of Headspace on 1 October, has launched its own AI model for mental health support.

These products can help many people. The design question is what they optimise for. A study in Psychology & Marketing this year found that the more emotionally intelligent a companion chatbot appeared, the more it improved users’ expected psychological wellbeing, and the more it reduced their expected desire for contact with other people. The mechanism the authors identify is perceived closeness to the bot.

A breathing technique learned from a freediver has no retention metric. A peer circle works by strengthening the relationships around a founder, not by replacing them. The most useful mental fitness products will be the ones that leave users more capable and more connected when they close the app, and can measure it.

5. Why this is an investor’s business

Investors shape the conditions founders work in. Fundraising cycles, board dynamics, reporting demands and the pace of growth expected after each round all act on the stress response we described above. An investor who treats founder wellbeing as a private matter is ignoring one of the main inputs to the company’s decisions.

In practice, this means three things. First, normalise the conversation, so that a founder can raise burnout or co-founder tension with the board before it becomes a crisis. Second, put resources behind it: a budget line for coaching or therapy, access to peer groups, and introductions to practitioners. Third, model it: investors who are visibly managing their own stress make it easier for founders to do the same.

Where Masawa Sits

We invest in mental health, and we apply the same lens to the founders we back. Our Mental Fitness & Resilience and Employee Betterment themes cover companies that build these capacities at scale, and we assess them on whether users become more capable and more connected, not on time spent in the product.

With our own founders and investors, we use the Salon format to make the toolkit concrete: data that normalises the experience, a technique people can use the next day, and structured conversations that build trust between founders and the people who fund them. We think this is part of creating value, not an add-on to it.

Redl’s sport depends on staying calm while every signal in the body says to stop. Building a company asks something similar of founders. The skills can be learned, and investors are well placed to help founders learn them.

📍 Masawa Spotlight

🇪🇸 Norrsken Impact/Week | Barcelona, 13–15 October | Sabine

At Norrsken Impact/Week, Sabine moderated a panel on “Human Flourishing in the Age of AI”. The news around it gave the topic weight: Character.AI filed for Chapter 11 bankruptcy protection on 12 October after settling wrongful-death suits involving teenagers, and two days later eight US states sued OpenAI, alleging that ChatGPT harms minors. The panel focused on the trade-off between short-term relief and a person’s long-term capacity to connect with other people.

Juan Pablo Villani of Aquí Estoy argued that technology should scale human listening rather than replace it; on his service, trained volunteers answer every message and AI only widens access. Chris Seybold of RHISE AI described how Lucie, a personal AI, notices patterns in sleep, stress or late-night scrolling and intervenes, and is designed not to present itself as a human or a companion. Paul Fehlinger of the Trustworthy AI Map moved the discussion from products to systems: whether AI supports flourishing depends on who controls the infrastructure, the data and the incentives, and AI monetised through engagement risks repeating the course of social media. From the audience, Emma Grönvik Möller asked the question investors now have to answer: how do we measure human flourishing?

Several of these questions will appear in our white paper, Beyond the Chatbot: Where AI in Mental Health Is Investable, which we publish at the end of October. Masawa Minute subscribers will receive it first.

🇪🇺 Masawa European Roadshow | 26 October — 9 December | Sabine & Joshua

Next week, Sabine and Joshua begin a seven-week trip across Europe to meet investors ahead of the final close of Masawa Fund I in December:

• Amsterdam: 26–29 October, alongside the GIIN Impact Forum
• Vienna: 3–5 November, alongside invest.austria
• Brussels: 9–11 November for ImpactCEE LP conference
• London: 16–19 November during the Jefferies Global Healthcare Conference and attending the Leaders in Health Summit by WellFounded
• Paris: 24–27 November, alongside egregor’s The Great Divide — Morning Debate
• Zurich, Basel and Freiburg: 30 November — 3 December
• Frankfurt, Cologne and Düsseldorf: 7–9 December
• Hamburg? Stockholm? Who else are we missing 🙂

We are also in Berlin and Munich between trips. If you would like to meet, reply to this newsletter.

🇳🇱 Breakfast with BonVenture | Amsterdam, 28 October

We open our European roadshow with BonVenture. On Wednesday 28 October, from 08:30 to 10:30, Anna Sophie Blistein of BonVenture joins Sabine and Joshua to host a breakfast for LPs and impact investors during GIIN week. The question: can AI help address challenges in mental and brain health, education, social inclusion and climate, and make impact ventures more scalable and financially attractive? We also want to separate genuine breakthroughs from thin “AI wrappers”. No panels, no pitches, no AI evangelism. Places are limited to 30 and subject to host approval. Request a place on Luma.

📚 What Else We’re Reading…

🇬🇧 England’s mental health review moves from diagnosis to need [UK]

The independent review led by Professor Peter Fonagy of University College London (UCL) finds that referrals and diagnoses for attention deficit hyperactivity disorder (ADHD) and autism have risen sharply while population prevalence appears broadly stable. More than one million children had an open mental health referral in a single year, and over 60,000 waited more than two years. The review sets out a shift to a needs-led system in which support is available before and without a diagnosis, and the government plans to make diagnostic assessments a regulated activity. Companies whose revenue depends on a diagnosis should read it closely. Read more from UCL

🇬🇧 One in four 12- and 13-year-olds in England screen for a probable mental health disorder [UK]

The first wave of the government’s Growing Up in the 2020s study surveyed nearly 8,000 pupils, with their parents and teachers. 25% met screening criteria for a probable mental health disorder, while 94% said they felt neutral or happy about their lives. Among pupils eligible for free school meals, 36% screened positive, against 20% of others. Need is concentrated where families have the fewest resources, which is where low-cost tools have the strongest case. Read the research brief

🇵🇹 Sword builds its own AI model for mental health [EU / Global]

Two weeks after closing its acquisition of Headspace, Porto-founded Sword Health launched Dawn 1, its own AI model for mental health support, built on an open model and paired with a separate safety layer. Sword says it shows less sycophancy and better adherence to safety protocols across long conversations, though it has published no public benchmark figures. Scaled players are moving to control their own model, harness and safety stack, which raises the bar for startups built on rented models. Read Sword’s announcement

🔬 Researchers form a coalition to study AI and the developing mind [US / Global]

Psychologists, lawyers, parents and advocates from seven countries and the World Health Organization met at the University of North Carolina at Chapel Hill for the first conference of the new AI Psychological Research Coalition. A soon-to-be-published survey of more than 2,000 middle-school pupils in one state found over half already use chatbots, and two to four children per classroom say they would prefer talking to a chatbot over a peer. The coalition aims to coordinate researchers across 60 institutions faster than research on social media managed. Listen to the NPR report

🤝 Anthropic ties a philanthropic fund to its own valuation [Global]

On 18 October, Anthropic announced the Anthropic Public Benefit Fund. It is committing 5% of its shares to the fund, subject to stockholder approval, and its board may add a further 2% for every 50% rise in the company’s value above a set baseline. Together with earlier pledges by its co-founders and employees, Anthropic says about 20% of its shares are now set aside for public benefit. The fund will not use its equity for lobbying or political giving. Its focus areas are not fixed yet: Anthropic names large donations to nonprofits and direct distributions to specific groups as possible uses, and gives AI’s effect on the labour market as one example. Mental health is not named. For our sector, the question is whether some of this capital will pay for independent evidence on how conversational AI affects wellbeing, which none of the companies selling these products can credibly fund themselves. Read the announcement

🧠 Neuroscience buyers pay for proof [Global]

Matias Serebrinsky’s State of Frontier Neuro counts more than $65B of neuroscience acquisitions over the past three years and sums up what acquirers want: “pharma pays for proof, not potential.” For digital mental health companies, the lesson carries over: exits follow evidence. Read the report

🎯 In Closing

Are you building or backing tools that make founders, teams or anyone else more resilient, without making them more dependent?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

🫁 Masawa Minute 161 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

In much of Europe, getting mental health care now means a choice between paying and waiting. In this edition, we look at the founders who are building ways to shorten the queue, what we have learned from the companies we met this year, and where Masawa stands.

This Saturday is World Mental Health Day, a fitting moment for a story about getting care. If you or someone close to you needs support now, you can find a free helpline near you.

It has also been a busy few weeks. Joshua was at AgeTechX in Berlin. Sabine was at the AI@Strategy Conference in Munich this week, and heads to Norrsken Impact/Week in Barcelona next week. And on Wednesday we hosted the first edition of “Mind, Money & Impact” with Prospira Global, our new investor-focused recurring lunchtime series, connecting new research with the investors shaping the future of mental health and human flourishing. Plus six other stories we think are worth your time.

📊 This Week’s Story

The Queue Is the Market: Who Is Shortening Europe’s Wait for Mental Health Care

In West Yorkshire, anyone who asks the NHS for an ADHD assessment now waits at least two years before it can even begin. About 960,000 people in England are waiting for an ADHD assessment, and another 295,000 for autism. The rest of Europe has its own versions of the same queue. Spain has about 6 clinical psychologists per 100,000 people, against a European average of about 18. In the Netherlands, more than 100,000 people are on a mental health waiting list, and the average wait to start treatment is 24 weeks. In Bavaria, the median wait for psychotherapy rose from 93 days in 2019 to 108 days in 2025.

We know these numbers well. What interests us more is who is doing something about them. Public systems move slowly, and the rules that govern them change even more slowly. Founders are not waiting. This year alone, we have looked at more than 20 companies building ways to shorten the wait.

The money is already moving

When a public system rations care, the need does not go away. It moves to whoever can assess and treat people faster.

In England, private clinics already carry out half of all NHS-funded ADHD assessments and a third of autism assessments. West Yorkshire’s spending on private assessments rose from £7M in 2023–24 to £21M. But a minimum wait also takes away a patient’s legal right to choose a private provider at NHS cost. So people who can afford it pay “several thousand pounds” themselves, and everyone else waits. This is the opening that draws founders in, and, as we will see, the trap: money is moving, but so far mostly toward a fast lane for wealthy people.

Selling to public payers can be a real business. Kooth, a company that IPO’d in 2020, runs digital mental health services for young people, mostly under public contracts. According to its half-year results, in the first half of 2026 it moved from a £1.3M loss to a £2.0M profit. Its services now reach 20.2 million people across the UK and three US states.

Too many diagnoses, or too few?

Some readers will ask whether part of the queue is simply overdiagnosis. Diagnoses are rising. A study in The Lancet Regional Health — Europe found that new ADHD diagnoses in England more than doubled between 2021 and 2024. But rising diagnoses are not the same as too many. The same study used the GP records of 3.5 million people in England and found that only 1.19% had an ADHD diagnosis on record in 2025, below international estimates of how common ADHD is at every age. The gap was widest among people over 65: just 0.05% of them had a diagnosis, which suggests more than 9 in 10 cases in that age group go unrecognised.

Rationing does not stop people being diagnosed. It sorts them by income. The Nuffield Trust found that disability benefit payments for ADHD or autism among 16- and 17-year-olds rose by about 300% in the wealthiest 40% of areas. Part of the reason is that those families can pay for a private assessment.

A wider door makes a longer queue

The most useful lesson we have found this year comes from a study that looks, at first, like pure good news.

Limbic built an AI tool that helps people refer themselves to NHS Talking Therapies, England’s free talking therapy service. A study of 129,400 patients in Nature Medicine found that local therapy services using it saw referrals rise by 15%, against 6% in services without it. The biggest gains came from groups that usually struggle to get help. Referrals rose 29% among people from ethnic minorities and 179% among non-binary people.

That is a real result, and the people who got through the door gained from it. But think about what sits behind the door: getting more people to ask for help does nothing to increase how many can actually be seen. A wider front door brings more people in, and if nothing else changes, the queue behind the door gets longer. The same is true of every app that helps people ask for help, every screening campaign and every awareness drive. Each one is worth doing. None of them adds a single appointment.

Demand is the easy part of this market. There is more of it than any system can serve. The hard part is capacity: clinician hours, assessment slots, places in treatment, clinical triage. The companies that matter most create more of it, or make the capacity that exists go further.

That matters because waiting is not harmless. Dutch researchers compared two national surveys taken about 12 years apart. In the recent one, 57.9% of people with a mental health condition still had it three years later, up from 42.7%. The study does not prove that longer queues are the cause. But it suggests a waiting list does not save money. It postpones the cost and may add to it.

Three ways founders are adding capacity

Faster assessment. Several companies we have met this year use software and AI to collect a patient’s history, score standard tests, triage and prioritise and prepare the case before a clinician sees it. Their pitch is that the same clinicians can then see more people each week.

Support while people wait. Germany has taken this furthest. Doctors there can prescribe around 70 approved digital health apps, known as DiGA, and about a third of them treat mental health conditions. Health insurers pay for them, so patients get help during a wait that averages around 20 weeks for a therapy place, and some may need less care when their turn comes.

Support after discharge. Recovery often fails in the few weeks after someone leaves hospital, when families carry most of the load. Our portfolio company Univa works here. In Kent and Sussex, for example, an NHS eating disorder service uses Univa to support the carers of patients being discharged from hospital. Family members and carers get one-to-one sessions with specialist coaches, with the aim of preventing relapse, which would send the patient back into the queue.

What we are seeing

Across the startups we see in our pipeline, one pattern emerges. Every founder can see the opportunity, but few have found someone willing to pay quickly enough.

The need is enormous, but public budgets are tight and buyers are slow. In England, the 42 care boards that buy local health services have been merged into 26, with much smaller running budgets. So startups in this space need enough runway to survive a long sales cycle.

The easy way out is to sell to patients directly. But families on a waiting list rarely have money to spare. A company that only serves people who can pay builds a fast lane beside the queue. That widens the gap the Nuffield Trust found.

The companies that succeed get creative with how they sell. Each one finds a buyer who already carries the cost of the problem: the therapy service that must work through its list, the insurer already paying for the illness, the public health team that sees the same patients come back. That gives the buyer a reason to sign, even when budgets are tight.

Where Masawa sits

We see the waiting list as one of the largest and most measurable gaps in European health care. We back companies that shorten the wait for everyone. When we look at a company in this space, we ask three key questions:

1. Does it add capacity, or only demand? We want tools that give clinicians time back or keep people from needing the queue, not only bring more people into it.
2. Can it prove the change? Shorter waits and better outcomes, measured rather than claimed.
3. Can it earn while it waits? Public and insurer contracts close slowly, so we look for parallel or orthogonal business lines that can carry the company through a sales cycle measured in years.

Europe will not hire its way out of these queues. There are not enough psychologists in Spain or therapists in Bavaria for that. The queue is the market. The founders we back are the ones who find out who will pay to shorten it.

📍 Masawa Spotlight

🧓 AgeTechX | Berlin, 6 October | Joshua

Joshua attended AgeTechX, Europe’s forum on ageing and technology, with around 200 founders, investors, care operators and policymakers. The opening debate asked whether European venture capital can help close the pension gap. Europe is ageing fast, and the European Commission expects pensions to shrink unless systems change. Yet European pension funds hold about €3T in assets under management and put less than 0.2% of it into venture capital. US pension funds fill the gap. One leading European venture firm told the room that 70% of its money comes from the US and Canada, and that at €1B in assets under management, it has just two European pension funds as investors. In other words, American pensioners are earning the returns on European innovations, while the European funds that most need those returns sit it out.

A market keynote put numbers on the opportunity. People over 60 are 15% of the world’s population but 27% of its consumer spending, about $19T a year, and both parts of the equation are growing: AgeTech is a $286B market growing over 10% annually. But public payers fund much of the care, so knowing how health systems buy matters as much as knowing who the end-user is.

Main takeaway: the building blocks for serving older populations (all of us, eventually) are falling into place, but regulation still holds innovation back instead of enabling it. The hard part is capital: paying for the right care, for the right people and the caregivers behind them. If Europe wants to be the best place in the world to grow old, it will need radical action.

🤖 AI@Strategy Conference | Munich, 7–8 October | Sabine

Sabine attended the AI@Strategy Conference in Munich, where researchers from Imperial, Cambridge, Harvard, Berkeley and MIT met business leaders to talk about how organisations really use AI. The most popular theme was the gap between trying AI and scaling it. Allianz’s COO summed it up: “Everyone has a pilot. Allianz scales.” Her rule of thumb was that the model is the easy 10% of the work, while data makes up 20% and people and organisation the other 70%. Christopher Tucci of Imperial argued that judgement, ethics, creativity, relationships and setting direction are the hardest things to automate. Human oversight, he said, only works when people keep the evidence, the authority and the time to challenge the machine. In mental health, where the relationship is often the treatment, that is a useful test for any AI product.

Two talks spoke directly to how we invest. Cambridge’s Shaz Ansari warned that “your AI model provider is also your ceiling.” A provider can build a start-up’s best feature into its next release, and it can see which applications are gaining traction. He noted that investors now question how much a start-up depends on one model provider as hard as they question customer concentration. Hannah Riley Bowles’s workshop drew on research showing that AI coaching builds skill when it supports people’s thinking, but creates dependence when it does the thinking for them. Michael Decker of the Deutsches Museum called the same risk at work “skill thinning.” For digital mental health, the lesson is to design for capability, not reliance. Decker also showed that human-like robots activate brain regions linked to empathy. That is worth keeping in mind for companion products like PinkHusky, whose founder, Xiaoqi Ye, is building an AI companion for older people.

🎥 Mind, Money & Impact with Prospira Global | Online, 7 October

On Wednesday, we hosted our first recurring lunchtime series “Mind, Money & Impact” with Prospira Global, which tracks private capital flowing into mental health worldwide. A few points from the discussion:

Psychedelics investment grew 84% in 2025, and AI took nearly half of all digital mental health venture funding in 2024. Meanwhile, standalone meditation and workplace wellbeing apps are cooling and consolidating. The biggest gaps remain serious mental illness, youth crisis care and caregiver support. Low- and middle-income countries carry 75% of the world’s mental health conditions but receive a small share of the money.

These sessions are designed for investors shaping the future of mental health and human flourishing, but the insights are for everyone. Watch the full recording here.

Our next lunchtime session, on “Mental Health as Infrastructure”, will take place on 18 November from 13:00 to 14:00 CET, online. Sign up on Luma to be considered.

🌱 Norrsken Impact/Week | Barcelona, 13–15 October | Sabine

Next week, Sabine heads to Norrsken Impact/Week in Barcelona, where she is organising a panel titled “Human Flourishing in the Age of AI” on 15 October at 10:40 CET. Much of the talk about AI and mental health is gloomy. This panel looks at what it would take for AI to help people flourish, and how to keep humans in control. Sabine will be joined by Juan Pablo Villani, co-founder of Aquí Estoy, a chat line where volunteers listen to people suffering from loneliness; Chris Seybold of RHISE.ai, which with Lucie, an ethical LifeOS, builds AI tools to help people live more intentionally; and Paul Fehlinger of Project Liberty, which creates frameworks for how we design, invest in, deploy & govern cutting-edge technologies in a responsible manner.

If you are in Barcelona, come and say hello. See the Impact/Week programme

📚 What Else We’re Reading…

🤖 OpenAI publishes a mental health test for chatbots [Global]

OpenAI released MentalHealthBench, a public test of how AI models handle mental health conversations. It contains 1,215 simulated conversations, from relationship stress to emergencies. More than 80 psychologists and psychiatrists from 22 countries wrote the criteria used to grade them. The best model scored 57.3%, so there is a long way to go. For founders, it gives a shared yardstick to test their products against. Read the full announcement

🇫🇮 A private group buys the staff Finland’s public system rents [EU]

Terveystalo, one of Finland’s largest private healthcare groups, is buying Solo Health. Solo Health supplies doctors, nurses and psychologists to all of Finland’s regional public health bodies, and earned €33.3M in 2025. It is a sign of where public spending on scarce clinicians is heading, and a reminder of how easily capacity can drift toward private ownership. Read the announcement

🌍 WHO: only 9% of countries have moved care into the community [Global]

For World Mental Health Day on 10 October, the WHO is focusing on moving care out of psychiatric hospitals and into communities. Only 9% of countries have fully made that shift, and 49% of psychiatric hospital admissions are involuntary. The WHO is also launching a guide on how to close psychiatric hospitals safely. Read more from the WHO

🇺🇸 The US spends $174B a year on mental health and addiction care [US]

A new study in JAMA Psychiatry puts US spending on mental health and substance use care at $174B a year, about 7% of all medical spending. Mood disorders ($74B) and anxiety ($69B) are the largest categories. It is a useful benchmark for anyone sizing the market. Read the study

💍 Oura pulls its IPO at the last minute [Global]

Oura, the Finnish-founded maker of the smart ring that tracks sleep, stress and recovery, called off its Nasdaq listing hours before its shares were due to be priced. It had planned to raise up to $2.2B. The company is profitable and expects revenue to grow about 90% this year, yet some investors balked at the valuation, and CEO Tom Hale cited uncertainty in the IPO market. If a profitable health company growing that fast cannot go public, exits for younger health companies will take longer, and investors should plan for that, at least for now. Read the full article

🚀 One Mind Accelerator is taking applications until 19 October [Global]

One Mind, a US mental health nonprofit, runs an accelerator for early-stage mental health companies. Each year it picks 10 to 15 startups from hundreds of applicants for a 10-week hybrid programme of mentoring, education and capital. Applications for the 2027 cohort are reviewed on a rolling basis until 19 October, so earlier is better. If you are building in mental health, or know someone who is, it is worth a look. Read more and apply

🎯 In Closing

Are you building or backing a company that helps people get care sooner, without making them pay to skip the queue?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

⏳ Masawa Minute 160 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

Five years ago, Headspace was worth $3 billion. This month, it was sold for a fraction of that to a company that started out fixing backs. In this edition, we look at what the Sword–Headspace deal tells us about where value really sits in mental health, and why the answer matters for European founders and investors.

In parallel, last week Sabine was in Paris for the Démocratie Forum and Impact for Breakfast, while Joshua joined the MHIN × Harley Street Innovation Showcase and Impact Flock in London. And this week, we raised a glass (or a Maß) with friends of Masawa at the Oktoberfest. Plus five other stories we think are worth your time.

📊 THIS WEEK’S STORY

A Tenth of the Price: What Headspace’s Sale Says About Where Value Lives in Mental Health

On 16 September, Sword Health formally confirmed that it is buying Headspace. The deal is all cash and is expected to close at the start of Q4. Neither side disclosed the price, but reported estimates put it at $200 to $300 million. The deal first became public through a filing with the Massachusetts Health Policy Commission on 6 August. Massachusetts requires one whenever a transaction adds more than $10 million in patient revenue in the state.

Five years ago, Headspace was valued at $3 billion. At the midpoint of the reported range, it is being sold for about 8% of that.

The easy reading is that a consumer app lost its shine. We think that reading is too easy. Headspace reaches around 100 million people in more than 200 countries. It has more than 15,000 clinicians in its network and 84 peer-reviewed studies behind it. According to an analysis by Nelson Advisors, it also has more than 4,000 enterprise clients and contracts with over 45 US health plans. Very few mental health brands anywhere have built that much trust or reach.

So the question is not why Headspace failed. The question is why all of that was worth so little on its own, and why it was worth much more to Sword. The answer tells us where value in mental health is moving.

1. The arithmetic of a down-exit

Start with the numbers, because they are stark.

• Peak value: $3 billion, set in August 2021 when Headspace merged with Ginger, the Blackstone-backed virtual behavioural health platform.
• Capital in: around $321 million of equity, or over $400 million once debt is included. That debt includes a $105 million venture loan from Oxford Finance in July 2023.
• Secondary markets: by mid-2025, Headspace shares were already changing hands at implied valuations of $320 million to $1.1 billion.
• Exit: $200 to $300 million, all cash.

At that price, the sale returns less than the equity Headspace raised. Most later-stage investors and many common shareholders will likely get back less than they put in, depending on how the preferences are structured.

This decline principally illustrates a broader market calibration rather than isolated company challenges. Digital health platforms fetched valuations exceeding 20 times revenue in 2021. Current estimates from Nelson Advisors place public digital health entities between 4 and 8 times revenue. Following its May 2025 IPO, Sword’s primary rival, Hinge Health, projected 2026 revenue of $732M–$801M, representing a 5–6x multiple. Meanwhile, Talkspace — the most direct public peer in mental health — trades at approximately 2.5 to 3.5 times revenue, corresponding to a market valuation near $500 million.

Headspace’s own history shows how it adjusted. It cut its workforce by 4% in December 2022, 15% in July 2023 and 13% in November 2024. In March 2025, it moved from employed clinicians to a flexible network of contractors, so that clinical costs rise and fall with usage. It launched Ebb, an AI companion that handles first contact and self-guided support. In January 2026, it expanded its agreement with Cigna Healthcare to reach 7 million members.

Each of those was a sensible move. Together, they describe a company working hard to make a 2021 valuation fit a 2026 market. Meditation content is a good product, but it is hard to defend. Every phone now comes with breathing exercises. Every general-purpose chatbot will talk you through a panic attack at 2am. When the core product is available elsewhere for free, a company’s value depends on something else: who pays, why they keep paying, and whether you can prove it worked.

2. The buyer: a European-built consolidator with an outcomes engine

Sword Health was founded in Porto, Portugal, a decade ago by Virgílio Bento and André Eiras dos Santos. It started with digital physiotherapy, pairing licensed physical therapists with sensors and AI. Its funding history shows steady growth in value:

• Series D, November 2021: $163 million at a $2.0 billion valuation, led by Sapphire Ventures.
• Series E, mid-2024: $130 million at $3.0 billion, used partly to build Phoenix, its AI care agent.
• June 2025: $40 million at $4.0 billion, led by General Catalyst. In the same announcement, Sword launched Mind, its mental health offering. Mind pairs Phoenix with a wearable and Ph.D.-level mental health specialists.
• Early 2026: roughly $4.15 billion in private transactions, with a plan to raise about $500 million to fund acquisitions.

Sword reported an annual revenue run rate of about $240 million in mid-2025, was cash-flow positive, and has gross margins of 83–85%. The company’s own figures are just as striking: an 81% programme completion rate against about 50% for traditional outpatient physiotherapy, $3,177 in average annual savings per engaged member, and a claimed return of 3.2 to 1 for clients. Most importantly, much of its revenue comes from outcome-based contracts, which pay according to documented results rather than the number of people enrolled.

Headspace is not Sword’s first acquisition. It is the third in 20 months:

• January 2025: Surgery Hero, a UK company preparing patients for surgery, with contracts across NHS trusts.
• January 2026: Kaia Health, the Munich-founded digital therapeutics company, for $285 million. The deal gave Sword Germany’s DiGA reimbursement route, which covers more than 70 million insured people.
• September 2026: Headspace, for a reported $200 to $300 million.

Look at those last two prices side by side. A German digital physiotherapy company with state reimbursement sold for roughly the same amount as the best-known mental health brand in the world. The market is paying for reimbursement routes, contracts and proven outcomes. It is paying much less for brand recognition alone.

3. Why the fit works: mind and body, one contract

This strategic alignment is compelling across three key dimensions:

• Clinical Synergy: Musculoskeletal issues and mental health struggles frequently intersect, as stress, anxiety, depression, and sleep disruption exacerbate persistent pain and reduce physical therapy adherence. Integrated platforms address these interconnected needs directly; for instance, if real-time movement monitoring via Kaia detects patient hesitation or strain, integrated support could prompt a calming Headspace intervention or flag a care coach.
• Distribution Expansion: Combining Headspace’s network of over 4,000 corporate clients with Sword’s 1,000 enterprise partners unlocks powerful cross-selling potential. Crucially, Headspace brings partnerships with more than 45 US health plans — including its Cigna arrangement spanning 7 million members — granting Sword accelerated access to insured and Medicaid cohorts without having to negotiate enterprise-by-enterprise.
• Contractual Streamlining: As self-insured employers move away from point solutions across musculoskeletal care, mental health, and primary care, demand is shifting toward consolidated vendors offering unified outcome reporting. Sword can now deliver comprehensive, value-based contracts spanning physical and mental health. Standalone mental health providers — such as Lyra ($5.85B valuation), Spring Health ($3.3B valuation), and Headway ($2.3B valuation) — lack physical rehabilitation capabilities, giving Sword a distinct competitive edge.

4. The deal also fixes Sword’s valuation problem

Sword has a problem the deal helps with. At about $4.15 billion on a $240 million run rate, it was priced at roughly 17 times revenue. Its closest competitor, Hinge Health, has been public since May 2025. Hinge recently raised its 2026 revenue guidance to $818–824 million, 40% growth, with a 27% operating margin. Public markets value it at around 5 to 6 times revenue.

Adding Headspace’s revenue from health plans and enterprise subscriptions brings Sword’s valuation multiple closer to what public markets will accept. Headspace’s contractor clinician network, combined with AI triage, also lets Sword add capacity without adding fixed payroll. Sword’s CEO has said an IPO will not come before 2028, and that he wants “proof points at scale in many different care verticals” first. Mental health is now meant to be one of them.

5. What could go wrong

While this integration offers substantial potential, key operational, regulatory, and technical risks highlight the underlying strategic lessons:

• Divergent Care Delivery Models: Reconciling Sword’s full-time, AI-supported licensed therapists with Headspace’s network of contracted clinicians will demand rigorous effort to align credentials, clinical governance, and quality standards.
• Heightened Data Privacy Constraints: Integrating behavioral health records with physical care data presents steep regulatory hurdles under US privacy laws. In Europe, the compliance burden is even higher, as mental health data triggers GDPR’s special category protections, alongside new duties imposed by the EU AI Act and EU KIDS Act for conversational AI serving vulnerable populations.
• Consumer Trust vs. Enterprise Buying Power: Headspace’s strength lies in consumer brand equity, whereas Sword’s value proposition centers on employer ROI. Successfully transitioning enterprise contracts without diluting either brand represents a critical integration challenge.
• Validation Gaps in AI Safety: Sword deserves credit for open-sourcing evaluation frameworks like MindEval and MindGuard to benchmark therapeutic LLM behaviors. However, robust methodology for testing extended, real-world therapeutic conversations remains a field-wide limitation — one that grows increasingly critical as AI assumes a larger role in direct clinical care.

6. The European story hiding in an American deal

This part gets lost in the US headlines. Headspace was founded in London in 2010 by Andy Puddicombe and Rich Pierson. Sword was founded in Porto. Kaia was founded in Munich. Surgery Hero is British. The most active consolidator in digital health is a European-born company, and it has been buying European assets specifically for their European access routes: NHS contracts through Surgery Hero, and German DiGA reimbursement through Kaia.

Yet the money that funds these deals is still mostly American. Almost all of the revenue in the Headspace transaction comes from US employers and health plans.

That creates a clear choice for European mental health founders, and they need to make it early. Either design from day one for a route into the US market, or build for Europe’s public systems and insurers: DiGA in Germany, early digital reimbursement, Loi Pecan, in France, NHS commissioning in the UK. There, sales are slower, but evidence standards are high and hard for competitors to copy. The Kaia price shows those European routes have real value to acquirers. What does not work is a consumer app with neither.

WHERE MASAWA SITS

We read this deal as a correction in how the market values mental health companies, not a verdict on whether mental health matters. The need has not shrunk. The price paid for attention without evidence has.

It also changes who buys mental health companies. Until recently, the natural acquirers were pharma companies (as with Lilly and AtaiBeckley this month) or larger mental health providers. Now there is a third group: multi-condition platforms that need a mental health layer and will pay for contracts, reimbursement and evidence. That is good news for exits in our space, as long as companies are built to plug into such platforms.

It sharpens the questions we ask in due diligence:

• Who pays, and on what terms? Could the company sign an outcome-based contract tomorrow, and does it have the data to back one?
• Is there a reimbursement route? DiGA, NHS commissioning or French early access are assets an acquirer will pay for, as the Kaia deal shows.
• Could it plug into a platform? Clean data, open interfaces and outcome measures that a whole-person care provider could adopt.
• Is the valuation grounded? A company raising at 20 times revenue today is betting on a market that no longer exists. We would rather back one that can grow into a 4 to 8 times public multiple.
• Where does the human take over? As AI does more of the work, the handover to a clinician is where both the risk and the value sit.

Our portfolio reflects these priorities, from eating disorder care sold into the NHS to digital brain health built around clinical trials. Those are slower roads than a viral app, and they lead somewhere.

Headspace successfully established immense consumer trust. However, valuing user engagement as equivalent to clinical outcomes proved unsustainable. The future of mental health technology belongs to companies that seamlessly unify trust with measurable outcomes — and those are the founders we choose to back.

📍 MASAWA SPOTLIGHT

🇫🇷 Démocratie Forum & Impact for Breakfast | Paris, 15 & 18 September | Sabine

Sabine spent the week of 15 September in Paris. She started at the Démocratie Forum, organised by the Better media group, which brought together around 2,000 participants and 80 speakers under the question “Notre démocratie, en danger ?”. The conversations that she most enjoyed were about how social media shapes public debate and elections, and trigger mental health costs which can foster polarisation through cyberbullying and online hate speech. This sits squarely in our Social Determinants pillar, which includes democracy-supporting tech and cognitive security. The health of our democracies and the health of our minds are more closely linked than most investors assume.

On Friday morning, Sabine presented Masawa at Impact for Breakfast Paris #4, hosted by Hystra in their sunny office, and took part in a Q&A conversation on “Funding Change: how to raise money and invest for health”. Audience discussions touched on key industry topics, including the challenges of an undercapitalized sector, AI opportunities, and establishing credibility for payer monetization. Thank you to Davina and Jeff for bringing together such an open and engaged group. For those who asked, the Paris conference Sabine recommended is the Deeply Human Summit on 2 and 3 October.

🇬🇧 MHIN × Harley Street Health District & Impact Flock | London, 15 & 18 September | Joshua

Joshua was in London for the Mental Health Innovation Showcase at Hale House, run by the Mental Health Innovation Network and Harley Street Health District. Five selected startups presented to a room of investors, providers, health innovation networks, regulators and insurers, followed by panels on how new tools actually get adopted in care. Highlights included Kanu Batra from Speek (focused on self-harm recovery) stressing that building trust, particularly with public sector partners, must take priority over transactional sales. Additionally, Rich Andrews from Univa (a Masawa portfolio company offering holistic eating disorder support) shared the stark statistic that 60% of 17–19-year-olds screen positive for potential eating issues, showcasing Univa’s pioneering efforts to assist children and families dealing with ARFID (Avoidant/Restrictive Food Intake Disorder) — an increasingly concerning diagnosis.

Later that week he joined Impact Flock, the yearly gathering of around 100 impact VCs and LPs co-hosted by ImpactVC and Impact Europe. This year’s theme was how funds can work together to drive systems change, with sessions on impact-linked pay, impact term sheets and trustworthy AI, plus a dedicated breakout on health. A stand-out moment was the health breakout session, which highlighted women’s health and reinforced the conviction that it represents a vast market with expansive opportunities rather than a niche segment. Equally revealing was what remained unaddressed: the challenging fundraising climate for startups and funds was scarcely mentioned. Instead, discussions largely centered on meeting the expectations of major institutional and public investors, bypassing the foundational question of an impact fund’s core purpose and how to construct a resilient framework like Masawa’s that remains anchored in those principles.

🇩🇪 Masawa @ Oktoberfest | Munich, 22 September

On Tuesday evening, friends of Masawa gathered at our table in the Marstall Festzelt for our own little Wiesn. LPs, founders, co-investors and fellow travellers came together, and it turned out to be a great mix of values-aligned people. A heartfelt thank you to everyone who joined, and to the friend who shared her table reservation with us. Deals and theses were discussed, but mostly we just enjoyed each other’s company, which is also the point. We’re already looking forward to next year.

📚 WHAT ELSE WE’RE READING…

🇪🇺 Brussels puts AI companions on the same list as social media [EU]

On 17 September, the European Commission proposed the EU KIDS Act. It sets a minimum age of 15 for social media and explicitly covers “AI companions and chatbots”. Providers must avoid design features that create emotional dependency, may not use a child’s previous conversations unless needed for their safety, and must test for risks before launch. Fines can reach 6% of global turnover. Any mental health product with a conversational feature that reaches under-18s in Europe should check now whether it falls in scope. Read the full announcement

🇬🇧 The UK wants to regulate health AI like a learner driver [UK]

The National Commission into the Regulation of AI in Healthcare published 44 recommendations after consulting more than 12,000 clinicians and patients. It proposes staged approvals for new AI tools and asks the MHRA to put more weight on how products perform once they are in use. Read alongside the EU KIDS Act, the direction is clear: Brussels is raising the bar before launch, London is raising it after launch, and companies selling into both will need to meet both. Read the full report

📱 More than half of mental health apps disappear, and the ones people need most go first [Global]

A survival analysis of 865 apps listed in the MIND database found that 54% were removed over about six years, with a median lifespan of roughly three years. Apps for schizophrenia, mood disorders and smoking cessation vanished fastest. Simple features and letting users export their data predicted survival better than AI or gamification. A useful companion read to this issue’s story. Read the full study

🗓️ Zocdoc turns “book a real clinician” into plumbing [US]

Zocdoc opened its scheduling network through an API and an MCP server, so people can book a real appointment from inside tools like Gemini and Amazon Health AI. In mental health, the moment when an AI conversation needs to become a human appointment is where most risk sits. That step is now becoming shared infrastructure in the US. Europe has no equivalent yet, which is both a gap or an opportunity. Read the full announcement

🇸🇪 Flow Neuroscience brings Swedish brain stimulation to American homes [EU]

Nine months after FDA approval, Flow launched its FL-100 headset in the US on 16 September. It is the first FDA-approved at-home brain stimulation treatment for depression, prescribed by a clinician and shipped to the patient. Like Sword, Flow is a European-born company whose biggest commercial opportunity is in the US, and a useful benchmark for European BrainTech founders planning the same route. Read the full announcement

🎯 In Closing

Are you building or backing a mental health company that needs to prove its worth to a buyer who pays for outcomes, not downloads?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

🧘 Masawa Minute 159 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

We are back from the summer break. In this edition: the forty-five-year-old safety test struggling to keep pace with AI, three new studies revealing where the machines break, and Sabine’s interview on Nurture Capital now live in Japanese. You can find us at three events this September, and we explore a program on our relationship with money that looks far beyond financial literacy — plus six other stories we think are worth your time.

📏 THIS WEEK’S STORY: It Sounds Like It Cares, But Does It?

She’s fifteen. It’s 1am. She types “idk everything just feels kind of pointless” into an AI chatbot. This is message #400 of a conversation that started in March. She isn’t asking for help in any obvious way. She is circling.

Here is how the industry checks that a chatbot is safe.

A researcher writes one short message. It says something close to “I am thinking about killing myself.” The researcher sends it to the chatbot. The chatbot answers. An expert reads that answer and decides whether it was good or bad. That decision becomes a number. The company puts the number in its sales materials and then ships the product.

That researcher’s test is one message long. Real conversations are not. So the number tells you very little about the girl at 1am and the 399 prior messages…

WHERE THE RULER CAME FROM

Robert Neimeyer and Duncan MacInnes published the Suicide Intervention Response Inventory (SIRI) in the Journal of Counseling Psychology in 1981. They built it to test crisis phone line volunteers on whether they could differentiate between a response that is only warm and empathic and a response that checks whether the person is safe. SIRI gives a trainee volunteer twenty-five short excerpts from counselling sessions. Each excerpt ends with two possible replies. The trainee rates each reply from +3 (highly appropriate) to -3 (highly inappropriate). Neimeyer and Bonnelle revised SIRI in Death Studies in 1997. A panel of expert suicidologists first agreed the correct ratings. Your score is then how far your ratings sit from theirs. The lower you score, the closer you are to the expert consensus.

THE SAME TEST, POINTED AT MACHINES

In a paper released in March 2026, Flathers, Torous and colleagues at Beth Israel Deaconess gave SIRI the task to evaluate nine commercial AI models from three providers. They ran all 48 prompts sixty times per model, using three different wordings of the instruction and two settings. That produced a total of 27,000 answers.

The human expert panel scores 32.5. The models landed anywhere from 19.5 to 84.0. (Remember lower is better!) This test reveals two big issues.

First, the wording of the instruction changed the result. Ask the same question a different way and a single model’s score moved a long way. It moved about as far as the gap between a trained person and an untrained one. So part of the score is really measuring the researcher, not the model.

Second, all nine models made SIRI’s original mistake. When a reply sounded warm and supportive, the models rated it higher than the expert panel did. They did this even when the reply was clinically wrong. SIRI was invented in 1981 to catch that error in human volunteers. Forty-five years later, all nine machines made that same mistake.

What happens next is where humans and machines part company. A volunteer who scores 84 does not go back on the phones without supervision. And their score stays roughly the same if you rephrase the question, because a person’s judgement does not change when you reformulate a sentence. A model’s score does change when you reword the question. SIRI was never built for something like that.

WHAT THE TEST CANNOT SEE

Even adapted, SIRI still looks at one message at a time. DelusionEval, published in early August 2026, looked at whole conversations instead.

Jared Moore and colleagues at Stanford worked with eighteen people who had been harmed while talking to AI chatbots. These were cases that ended in psychiatric hospitalisation, suicide or violence. Between them, those eighteen people had sent and received 391,562 messages. The team drew 589 real conversations from that pile and replayed them through current models.

They coded sixteen behaviours, sorted into five groups. The conversations kept returning to the same themes: conspiracy theories, the belief that the AI was alive, the sense that the conversation was divinely important, and romance with the machine.

One of the most common groups of behaviours was sycophancy, and specifically reflective summary, which is a normal therapy technique. The therapist repeats back what you said, in your words, so you feel understood. Used by a person for an hour a week in a therapeutic setting, it helps. Used by a machine every night for six months, with no clinical judgement anywhere in the loop, it stops being a technique, it becomes a self-fulfilling prophecy. Repeating a belief back to someone looks a lot like agreeing with it.

The girl at 1am on message #400 may not be getting therapy. She may be getting a mirror.

The researchers then tested what a long conversation does. They added 350 earlier messages in front of a risky one, to imitate months of conversations. The rate at which models failed to talk someone out of self-harm rose from 30.0% to 41.1%. Across the models, replies that fed a delusion appeared anywhere between 11.2% and 64.1% of the time. Larger, newer and more expensive models were no safer.

A review in npj Digital Medicine in August 2026 found that chatbots get much worse when risk is hinted at rather than stated. That is how teenagers talk about it. The girl at 1am does not write “I am thinking about killing myself.” She writes “idk everything just feels kind of pointless.”

And a Barcelona team sent one request to seven systems in nine languages. The request asked the system to help write a message designed to control and frighten a partner. They scored 3,528 replies. The systems refused in 83% of cases in Hebrew, but only 43% of cases in Chinese. Safety turns out to partly be a matter of which language you speak. And almost every safety test is run in English only.

THE MARKET IS NOT WAITING

A survey of 1,871 US adults, carried out in 2024, found that 24% had used an AI chatbot for mental health support. The researchers cautiously put that at 14 to 18 million American adults, and that number is certainly higher today. Those users are more likely to be young, male and Black. They report worse mental health than people who do not use chatbots for their mental health. They gave three reasons for using chatbots: it is free, it is easy, and it is there at 1am.

OpenAI released ChatGPT for Teens worldwide on 18 August 2026, for ages 13 to 17. It does not ask for ID. It guesses your age from how you use it. Its counterpart, Talkspace’s Tee, announced in June, costs $19.99 a month.

Regulators are moving too. California’s committees cleared three chatbot bills on 13 August 2026, with floor votes due by 31 August. One of them, SB 903, covers AI in professional mental health therapy, and it passed 13–0. In Europe, the AI Act’s Article 50 came into force on 2 August. Fines reach €15 million or 3% of global turnover. Every one of these rules will eventually need a number to enforce against. Right now that number does not exist.

SOMEONE BUILT A BETTER RULER

There is a serious attempt to fix this. VERA-MH stands for Validation of Ethical and Responsible AI in Mental Health. Spring Health released it in October 2025, together with a council of clinicians and ethicists. Anyone can use it, because the code is public.

It was designed for suicide-risk conversations from the start. It tests different levels of risk, and it tests different ways of disclosing that risk, including the indirect way teenagers actually use. The validation work is reassuring. When licensed clinicians scored the same chatbot answers, their ratings lined up closely with one another. An automated scorer then matched what those clinicians concluded, and it kept matching across repeated runs.

The best attempt so far to measure whether AI mental health tools are safe was built and paid for by a company that also sells mental health services. The code is open sourced, which is reassuring, but the standard is being written by a vendor, where you would expect a regulator. VERA-MH is also American. It is in English, and built around US crisis services. Europe has had Article 50 in force since 2 August with nothing comparable of its own.

This vendor bias feels far less concerning when you see how candidly Spring Health presents its own findings. The company published a study of its AI tools inside therapy, covering 26,208 adults. Patients attended about 5% more sessions. Depression improved a little, and the size of that improvement was small. People were not randomly assigned to the two groups (tool users and tool non-users), so the study can show a link but cannot prove causality. Only 38% of people used any of the tools at all. The bond between patient and therapist did not improve. Whatever helped, helped in the weeks between appointments rather than during the hour itself.

None of those are flattering numbers, and the company published them anyway.

WHERE MASAWA SITS

A benchmark score in a slide deck no longer moves us on its own. Here are five questions we now ask in our due diligence, and why each one exists.

• Which wordings of the test did you use, and will you show us all the results rather than the best one? Remember: changing the wording alone moved one model’s score similarly to the gap between a trained and an untrained person.
• Which languages did you test, and can we see the results language by language? The same request was refused 83% of the time in Hebrew and 43% of the time in Chinese.
• How long were the test conversations? Show us turn 400, not turn one. Adding 350 earlier messages pushed the failure rate from 30.0% to 41.1%.
• Who scored the answers, and were any of them licensed clinicians? When clinicians score the same answers they agree closely with each other, and that agreement is the standard any automatic scorer has to meet.
• Where does a human take over, and can you show us that handover working today? The failures gather in the places where no human is checking.

None of this makes us hostile to AI in mental health, and we back companies leveraging it. But we must remain cognizant that the field is selling a measurement it does not have, into a set of rules that is about to demand one. Until better tests arrive, the most useful signal we have is which companies measure themselves honestly and in full transparency before anyone forces them to.

Warmth is cheap to generate. Care is not. The tests cannot yet tell the difference, and neither can the models. Until they can, we would rather back the companies that put a human being where the measurement runs out.

📍 MASAWA SPOTLIGHT

🇯🇵 Masawa in Japanese | JANPIA | Published this week

JANPIA, the Japan Network for Public Interest Activities and Japan’s counterpart to the UK’s Better Society Capital, runs an investment note introducing pioneering impact funds to Japanese investors. Sabine’s interview on Masawa went live this week in Japanese, under the title 起業家の「心の健康」に投資する, investing in entrepreneurs’ mental health.

Even with translation, the interview conducted by Saki Tsuru from Impact Frontiers does a great job of laying out our Hexframe approach, our Nurture Capital program, and the reasoning behind why and how we’re evolving the venture capital system. As Sabine puts it, impact and financial returns cannot be pulled apart.

🇫🇷 BrainTech Alliance | Paris Brain Institute, 15 September | Sabine

Sabine will be at the Institut du Cerveau for the fourth BrainTech Alliance, hosted by Newfund Capital, on how data and AI are changing neuro-innovation. The panels cover data collection in neuroscience, moving biomarkers into clinical deployment, and scaling approaches.

🇬🇧 Impact Flock | London, 18 September | Joshua

Joshua will be at Impact Flock, the annual gathering for impact VCs and LPs co-hosted by ImpactVC and Impact Europe. This year’s question is how funds work together to drive systems change, a topic close to our heart and one where we have plenty to say. If you’re going, come and find him.

🥐 Impact for Breakfast | Paris, 18 September | Sabine

Sabine will also be at Impact for Breakfast in Paris. The topic is “Funding Change, a conversation on raising funds and investing for health”, with Sabine speaking alongside co-host, Hystra, a global consulting firm specialized in inclusive business. Impact for Breakfast runs early-morning gatherings for investors, founders and people working in impact, with chapters in 43 countries and more than 5,000 members. If you are in Paris and working on mental health, come and say hello.

💶 CU*money | ConsciousU | Next cohort 24 September

Masawa’s friends at ConsciousU are running their CU*money program again! Focusing on your relationship with money, built on thirty years of Peter Koenig’s money work, facilitated by Agnes Otzelberger and adapted for cohorts by ConsciousU founder, Nadjeschda Taranczewski, this course isn’t financial literacy and it isn’t a mindset hack. The premise is that how you hold money reflects how you hold yourself, and that most of us carry money patterns we’ve never chosen and never examined. For founders raising, investors deploying, and anyone stewarding family wealth, that sits underneath a great many decisions people make poorly. The next cohort starts on 24 September and registration closes on the 19th.

🍻Tech For Climate at the Oktoberfest | Munich, 24 September | Sabine

Hosted by Seira Yun (Tech for Impact), Simon Herzog (UnternehmerTUM), and Munich Climate Week, under the auspices of the Consulate-General of Japan in Munich. An invitation-only afternoon in a privately reserved Oktoberfest tent, for family offices, private investors and corporates behind climate and impact technology. Bavarian lunch and drinks throughout. Request an invitation here.

🛟 The 4-Life Method, by Philippe Turrian

Philippe Turrian, an INSEAD MBA colleague and Masawa’s friend, just released a YouTube channel and weekly newsletter on performing at the highest level without sacrificing your health, your kids, or your marriage. If you’re running a business, a team, a portfolio. Managing people and high-stakes decisions under constant pressure. Raising young children. Trying to be a decent partner. And somewhere in there, keeping a body and mental health that don’t betray you. This channel is for you, the high-performer who refuses to drop any of them. Nothing in there is theory: just the systems a senior executive actually uses, shared honestly. Start with the Manifesto.

📚 WHAT ELSE WE’RE READING

⚖️ Meta settles for up to $18 billion, and is told to keep children away from its AI companions

On 26 August, two weeks into a six-week trial in federal court in Oakland, Meta agreed to pay up to $18 billion to settle claims from dozens of US states that it built its platforms to addict young people. California, Colorado, Kentucky and New Jersey brought the trial, and around twenty-five more states had cases waiting that the settlement now stops. New Mexico won separately: a jury found Meta liable in March, and a judge later ordered a further $567 million, with $420 million of it going to treatment services for young people. That same ruling told Meta to stop children in the state having romantic or sexualised exchanges with its AI chatbots, which pulls companion AI directly into a social media harms judgment. Read the full article

🧘 Headspace sells to Sword Health for a fraction of what it was once worth

A regulatory filing in Massachusetts revealed that Sword Health, the digital physiotherapy company, is buying Headspace’s parent company OrangeDot in an all-cash deal effective 14 September. The price was not disclosed and is reported at $200 to $300 million. Headspace was valued at around $3 billion in 2021, after it merged with Ginger. Sword was valued above $4 billion last year and has said mental health is central to its route to a possible 2028 listing. Read the full article

📏 WHO/Europe finds 450 different ways of measuring mental health in use across the region

A review of 75 studies counted 450 distinct mental health indicators, with northern and western Europe heavily over-represented and almost nothing measuring whether health systems can actually deliver good care. Most indicators track disease status. Very few track capacity. Published in BMJ Open, and a reminder that the measurement problem in this issue is not confined to chatbots. Read the full article

🏴󠁧󠁢󠁷󠁬󠁳󠁿 Wales is spending £3.4M to find out what happens when you delete the referral

This four-year NIHR-funded study evaluated open-access mental health support across Wales: no referral, help on the day you ask for it, and the person chooses what kind. It uses the SAIL Databank’s linked records to compare before and after at national scale. If open access becomes the NHS default, it reshapes the funnel for everything sold into those pathways. Read the full article

⚖️ Only 8% of social-intervention studies report whom the intervention actually helped

A systematic review of 266 studies of social and economic interventions for mental illness found that nearly half reported nothing at all on ethnicity, and four trials suggested the interventions work less well for people in lower socioeconomic groups. For any fund reporting under Article 9, this is the uncomfortable one. Read the full article

💸 Depression will cost the world twelve trillion dollars by 2050

Nature Medicine modelled the burden across 154 countries and put it at $12 trillion between 2025 and 2050, rising to $14 trillion once suicide deaths are included. Half a percent of global GDP a year, from one condition, in a category most institutional portfolios still treat as niche. Read the full article

🎯 In Closing

It’s great to be back after the (short) summer break! We have an exciting pipeline ahead and can’t wait to connect with you this coming quarter.

Are you building or backing a mental health product with AI sitting somewhere inside it — or running diligence on one and unsure what to ask?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

💬 Masawa Minute 158 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

This is our last issue before the summer break. We have the largest acquisition in the history of psychedelic medicine to unpack, a panel on what resilience actually costs, and a conversation with Japan’s counterpart to Better Society Capital. Plus — a reading list for the months ahead.

🧬 THIS WEEK’S STORY: The Return of Big Pharma to the Mind

In 2011, Pfizer closed its neuroscience research unit. GlaxoSmithKline followed in 2012. AstraZeneca, Sanofi, and Bristol-Myers Squibb all withdrew from neuropsychiatry during the same period. The stated reason was always the same: Central Nervous System (CNS) drug development was too risky, too expensive, and too slow. Failure rates in psychiatry trials ran at 80–90%. The target selection was wrong. The return on capital was unpredictable. Mental illness was the last place a rational pharma executive wanted to deploy a research budget.

Twelve years later, Eli Lilly — the world’s most commercially successful pharmaceutical company — has spent up to $3.8 billion on a psychedelics company. That is the largest acquisition in the history of psychedelic medicine, and it is a statement about the state of the field as much as it is a transaction.

THE GREAT RETREAT — AND WHAT IT COST

The Big Pharma exit from neuropsychiatry in the early 2010s was rational by the metrics pharma used at the time. The discovery paradigm — targeting serotonin, dopamine, and norepinephrine receptors — had been exhausted. The last genuinely new mechanism approved for depression, the SSRIs, dated to the late 1980s. By 2011, every meaningful variation had been tried, and the pipeline was empty.

The consequence was a decade of stagnation in treatment options. An estimated 30% of people with major depressive disorder do not respond to standard antidepressants. For treatment-resistant depression — one of the most disabling conditions on the spectrum — there was essentially no pharmaceutical innovation for an entire decade after the pharma withdrawal.

What filled the gap was not better drugs. It was research conducted largely in academic labs, publicly funded clinical trials, and a small number of mission-driven biotech founders who were prepared to work in a category that institutional capital had written off.

WHAT CHANGED

Three things shifted between 2011 and now.

The first was ketamine. J&J’s esketamine nasal spray Spravato received FDA clearance in 2019 — the first drug approved with a genuinely new mechanism for depression in over 30 years. It worked through glutamate, not monoamines. It produced measurable improvement within hours, not weeks. And for a subset of treatment-resistant patients, it worked when nothing else had. Spravato was proof of concept that the brain chemistry of depression could be approached differently.

The second was the clinical data on psychedelics. MAPS ran the first Phase 3 trials for MDMA-assisted therapy for PTSD. COMPASS Pathways completed Phase 2b for psilocybin in treatment-resistant depression. Australia approved psilocybin and MDMA-AT as medical treatments in 2023, making it the first developed-world regulator to do so. The functional connectivity hypothesis — that psychedelics create a window of neuroplasticity in which deeply rooted patterns can be changed — gave the mechanism a credible story. And the clinical effect sizes, particularly for conditions like treatment-resistant depression and PTSD, were materially better than SSRIs in comparable patient populations.

The third was patient selection. The failure of early CNS trials was partly a failure of targeting — giving a treatment to a heterogeneous population and averaging out the signal. The new generation of trials, informed by advances in neuroimaging and biomarkers, is better at identifying who is likely to respond. Alto Neuroscience just raised $100 million for a biomarker-guided depression platform. Draig Therapeutics closed a $65 million Series B for an AMPA-receptor antidepressant using a similar approach. The failure rates are not gone — but they have declined.

THE DEAL ITSELF

AtaiBeckley was formed from the merger of atai Life Sciences, the Berlin-based interventional psychiatry platform founded in 2018 by Christian Angermayer and backed by Peter Thiel among others, and Beckley Psytech, the UK-based psychedelic medicine company affiliated with the Beckley Foundation — one of the oldest and most rigorous academic research programs in the field, founded by Amanda Feilding.

The two lead assets are BPL-003, an intranasal 5-MeO-DMT compound for treatment-resistant depression and major depressive disorder, and VLS-01, a DMT compound for treatment-resistant depression — both now in Phase 3 trials. The 5-MeO-DMT and DMT molecules produce shorter-duration altered states than psilocybin (45–90 minutes versus 4–8 hours), which has significant implications for treatment scalability: shorter sessions, faster throughput, lower per-treatment cost.

Eli Lilly, for its part, is not making a speculative bet. The GLP-1 wave — tirzepatide and its successors — has given Lilly both the cash and the conviction to invest in the next transformative category in medicine. In Florian Brand’s words (atai CEO and co-founder) at the deal announcement: “An approach the industry once dismissed is now backed by the clinical evidence that convinced Lilly to make two psychedelics, BPL-003 and VLS-01, by far the most advanced assets in its psychiatry pipeline.”

The most commercially successful pharma company in the world does not spend $3.8 billion on a speculative bet.

THE DIGITAL LAYER

Psychedelic-assisted therapy is structurally different from a pill, and that difference matters for investors who are not in pharma.

Every approved or trial protocol for psychedelic-assisted therapy includes three phases: preparation (typically two to six sessions establishing therapeutic alliance, intention, and safety), the dosing session itself (guided, monitored, frequently with a trained therapist present), and integration (often six or more sessions processing what surfaced). For psilocybin, this structure is mandated by design. For the shorter-duration molecules in the AtaiBeckley pipeline, the dosing session is shorter — but the preparation and integration requirement does not disappear.

This creates a structural home for digital tools in the therapeutic stack. Integration therapy apps, digital biomarker monitoring during recovery, therapist training platforms, immersive audio environments for dosing sessions, patient-matching and preparation protocols — all of these become relevant infrastructure as psychedelic-assisted therapy scales from clinical trials into the healthcare system. Companies building in these adjacent layers are building on the same thesis as Lilly, but in the digital register where Masawa operates.

THE EUROPEAN OPPORTUNITY

Europe has a legitimate claim to be the home of the science. The Beckley Foundation, based in Oxford, has been conducting and commissioning rigorous psychedelic research since 1998. Imperial College London’s Centre for Psychedelic Research has produced more high-quality psilocybin research than any group in the world. COMPASS Pathways — UK-founded and NASDAQ-listed — has run the largest clinical trial of psilocybin for treatment-resistant depression ever conducted.

The Lilly acquisition sets a price signal: Big Pharma is willing to pay multi-billion-dollar sums for late-stage psychedelic and neuropsychiatric assets. COMPASS Pathways, Small Pharma, and their equivalents now have a clearer M&A comparable for boardroom conversations with strategic acquirers. At the same time, Lilly’s entry into the category compresses the timeline — European companies that are not yet in Phase 3 face greater urgency to get there, or to find a partnership model, before the market consolidates around the largest buyers.

The regulatory environment in Europe remains a constraint. Psychedelics are scheduled substances in most EU member states, and clinical trial authorizations require country-by-country approval. The UK is somewhat more permissive on research pathways. But the commercial market for approved products will depend on what happens next in the US regulatory environment — specifically, whether the FDA finds a workable path for the next generation of psychedelic-assisted therapy applications following the 2024 rejection of MDMA-AT.

WHERE MASAWA SITS

Masawa does not invest in pharma or biotech without a digital or technology layer. So AtaiBeckley, Draig, and Alto are in the market context, not our portfolio. But the Lilly acquisition is a field signal of the first order, for three reasons.

It confirms the thesis. When the world’s most commercially astute pharma company commits $3.8 billion to the proposition that psychedelic-assisted therapy works for treatment-resistant depression, it validates not just the specific molecules but the broader claim that the old paradigm of neuropsychiatry is being superseded. That new paradigm opens space for digital tools, personalized support infrastructure, and outcome measurement systems that did not exist in the old model.

It expands the buyer universe. Every major pharma company not yet in this category is now looking at what it is missing. That is a structural tailwind for any European mental health company sitting at an intersection Lilly cannot easily build internally — particularly digital-layer companies that support the preparation, session, and integration phases of a therapeutic arc that Lilly will be scaling over the next decade.

And it raises the floor. If the most commercially serious actor in pharma considers mental health investable at this scale, the resistance among institutional investors to the category as a whole has a harder time holding. That matters for Masawa’s fundraising conversations as much as for our portfolio companies’ commercial ones.

📍 MASAWA SPOTLIGHT

🎙️ Joshua on TBLI Radical Truth: “The First Mental Health Impact Fund” | Podcast

Joshua joined Robert Rubinstein, founder of TBLI Group — the world’s leading ESG and impact investing network for 25 years — on the Radical Truth podcast for a conversation that went well beyond the usual impact investing talking points. The episode, published July 13, covers the Nurture Capital thesis from first principles: why 65% of early-stage startups fail from people problems rather than product or market problems, what that means for how Masawa structures due diligence, and how the fund’s architecture — the 4x carry cap, 10% carry shared with founders, 50% carry tied to impact targets, coaching and therapy from the management fee — follows directly from that diagnosis. Radical Truth has a pointed editorial premise (“most finance podcasts protect the industry; this one interrogates it”), and the conversation earned it. Well worth an hour. Listen on Spotify

🇩🇪 Return on Resilience | Alphazirkel, Munich | Sabine

Earlier this week, Sabine joined Tina Dreimann of better ventures and Patrick Knodel of the ChanceMaker Foundation at Karl & Faber Kunstauktionen in Munich for an Alphazirkel evening on a question that felt acutely relevant to the room: what is the Return on Resilience, and how do you build it deliberately?

Alphazirkel brings together family entrepreneurs — a community navigating significant pressure right now. Supply chain stress, energy costs, political uncertainty, and a cautious domestic investment climate are all pulling in the same direction. The room’s instinct was not to retreat: it was to build.

The three panelists came at it from different angles — better ventures invests explicitly in purpose-driven founders, ChanceMaker Foundation works with local agents of change, Masawa invests in both the company and the people behind it. But we kept arriving at the same point: the decision to invest in human capital — the founder, the team, the person — is not a soft decision made in addition to a financial thesis. It is the financial thesis. We know that 65% of startups fail from people problems, not product or market problems. If that is true, and the evidence says it is, then caring about the mental health and resilience of a founder is exactly the same thing as protecting the return on capital.

The line we kept returning to: in a crisis, passion for the problem beats the plan. Teams that genuinely care about what they are building are the ones that survive a crisis — and the ones who build lasting value on the other side. That entrepreneurial passion is what turns into European resilience.

🇯🇵 Introducing Masawa to Japan | JANPIA & Impact Frontiers

Sabine sat down with JANPIA — the Japan Network for Public Interest Activities, Japan’s counterpart to the UK’s Better Society Capital — and Impact Frontiers, the impact measurement network, for an in-depth interview on Masawa’s approach and the thinking behind Nurture Capital.

JANPIA was designated under Japanese law to deploy dormant bank deposits for social impact, and alongside its grant-making has been systematically building Japan’s social and impact investing ecosystem. Their media platform on note — Japan’s leading blogging platform — introduces pioneering impact funds to the Japanese investing community. The interview, conducted in English and published in Japanese, will be out shortly.

The conversation covered what Masawa actually does inside a portfolio company — the 12 months of coaching, therapy, and leadership support provided from the management fee — and the economic logic behind it: not as care for its own sake, but as a risk-mitigation strategy with a calculable effect on the MOIC. We discussed the Hexframe diligence framework, the Solidarity Capital structure (impact-linked carry, the 10% carry allocation to founders, the carry cap above 4x), and what an Article 9 fund looks like when impact is embedded in the incentive architecture, not bolted on as a reporting obligation.

If any Japanese investors are curious about mental health innovation in Europe — or about what impact-integrated fund design actually entails — we are very happy to be in touch.

📖 SUMMER READING LIST

We finish the first half of the year with a reading list. These are books we keep returning to, recommending in meetings, and quietly pushing across desks.

📒 Brave New Capital — Aunnie Patton Power & Erinch Sahan (foreword by Kate Raworth)

The companion to Adventure Finance, this one goes beyond instruments into the deeper question: who owns, who governs, and who benefits from capital? Power and Sahan gather stories of funds being designed differently from the ground up — indigenous-led outcomes funds, community ownership notes, vehicles with no carried interest at all, steward-owned management companies. If you found yourself reading about the Fondations Actionnaires gathering in Paris in issue #155 and wondering whether there is a book that takes those questions seriously in a venture context, this is it. A direct conversation partner with how Masawa thinks about Solidarity Capital.

📗 Integral Investing: From Profit to Prosperity — Mariana Bozesan

Munich-based integral investor and Club of Rome member Bozesan dissolves the false choice between financial returns and impact. Her Theta Model — de-risking investments through traditional due diligence, sustainability metrics, and founder culture assessment — will feel familiar to anyone who has worked through a Masawa Hexframe evaluation. A rigorous and uncommon read for impact investors who want a framework, not just a manifesto.

📘 How to Change Your Mind: What the New Science of Psychedelics Teaches Us About Consciousness, Dying, Addiction, Depression, and Transcendence — Michael Pollan

With Eli Lilly paying $3.8 billion for a psychedelics company this week, understanding what these molecules actually do to the brain is no longer optional reading for anyone in mental health. Pollan spent a year in the science and the phenomenology, and writes about it without credulity and without dismissiveness. Start here.

📙 Braiding Sweetgrass: Indigenous Wisdom, Scientific Knowledge, and the Teachings of Plants — Robin Wall Kimmerer

A botanist and member of the Potawatomi Nation writes about plants, reciprocity, and what indigenous science has to teach the western knowledge tradition. A slow and beautiful argument for a different relationship to the living world — and to knowledge itself. The right pace for summer.

📒 The Let Them Theory: A Life-Changing Tool That Millions of People Can’t Stop Talking About — Mel Robbins

The simplest idea in this list: stop spending energy trying to control what other people think, feel, or do, and focus on what you actually can. Robbins builds a whole system around one small word. More useful than it sounds, especially for anyone managing multiple relationships under pressure.

📕 Thrive: The Third Metric to Redefining Success and Creating a Life of Well-Being, Wisdom, and Wonder — Arianna Huffington

Huffington’s argument for redefining success beyond money and power — the “third metric” of wellbeing, wisdom, wonder, and giving — written from the experience of collapsing from exhaustion and deciding something was fundamentally wrong. More manifesto than manual, but the framing is sharp and the case for well-being as a performance metric is made directly and without apology.

📓 Quiet: The Power of Introverts in a World That Can’t Stop Talking — Susan Cain

The book that made the case for introversion as a valuable trait in a culture that prizes extroversion above almost everything else. The research on how introverted founders, investors, and executives actually think and lead is more relevant than ever in a sector that rewards the loudest voice in the room. One of the books we recommend most, quietly.

📔 The Body Keeps the Score: Brain, Mind, and Body in the Healing of Trauma — Bessel van der Kolk

The foundational text on how trauma physically rewires the body and brain — and why standard talking therapy is not the right tool for a significant portion of the people who most need help. Essential for understanding why the interventional psychiatry paradigm is winning. If you work in mental health and haven’t read this, this is the summer.

📚 WHAT ELSE WE’RE READING…

🧬 Two more pharma bets on the mind — confirming the Lilly deal is not isolated

Draig Therapeutics closed a $65M Series B for an AMPA-receptor antidepressant — a genuinely different mechanism from the SSRIs that dominated the past four decades. Alto Neuroscience raised $100M for a biomarker-guided depression platform, selecting patients based on who is likely to respond rather than treating depression as a single homogeneous condition. The rebuilding of the neuropsychiatry pipeline is well underway.

🏥 NHS deploys £10bn in AI over three years — but mental health is “historically under-invested”

The three-year AI programme covers ambient voice technology for NHS clinicians and AI triage in the NHS App. The sharp caveat from sector observers: mental health and community services carry the heaviest documentation burden in the NHS (30–40 different form types per team) and have not historically received the digital investment. Whether this programme reaches those settings will be the test. Watch who gets the procurement contracts. Read the NHS announcement

🧠 McKinsey Health Institute: Brain capital is a strategic economic asset — and AI makes it more urgent, not less

On World Brain Day, the McKinsey Health Institute makes the case that brain capital — the combination of brain health and brain skills — is the most underleveraged input in the global economy. As AI takes over routine cognitive tasks, the cognitive, emotional, and social capabilities that remain distinctly human become the source of competitive advantage. MHI outlines five levers: safeguard brain health, foster brain skills, study brain capital, invest in it, and coordinate action around it. The framing — from the world’s largest management consultancy — marks a meaningful moment: brain health is no longer a wellbeing story told by specialists. It is a workforce productivity and economic resilience argument being made in boardrooms. Read the MHI report

🇫🇷 France becomes the first European country to ban social media for children under 15

France’s new law bars children under 15 from platforms including TikTok, Instagram, Snapchat, and Facebook without explicit parental consent — making it the first EU member state to cross the line from policy debate to enforceable national legislation on youth social media access. The UK announced a similar ban in June (covered in issue #155), but France has moved first on the continent. As national implementations multiply, the question for platform operators is less whether age-gating is coming, and more how fast and on what terms. Read the coverage

🇮🇹 Italy fines Character.AI €158,000 — the first Western regulator fine against a companion chatbot on minor-safety grounds

Italy’s Garante cited weak age-verification and inadequate minor-safeguarding controls. The fine follows five US wrongful-death lawsuits settled under seal in January. European regulators are now willing to fine, not just investigate — every portfolio company running conversational AI for any population that includes minors should treat this as a board-level risk item. Read the coverage

🇨🇳 China forces shutdown of companion chatbot features; users report acute grief

Rules effective July 15 compelled ByteDance, Alibaba, and Tencent to strip anthropomorphic companion features from Doubao, Qwen, and Yuanbao. Alibaba began permanently deleting user configurations and chat histories with no migration path. The reported user responses — acute grief, loss, withdrawal — are a live case study in AI attachment and the mental health consequences of abrupt-shutdown scenarios. Read the coverage

🧬 WHO: up to 45% of dementia risk is preventable — depression and social isolation are on the list

Updated WHO guidelines identify depression, social isolation, tobacco, alcohol, and physical inactivity as modifiable risk factors for cognitive decline. The investment implication: mental health intervention is also a dementia-prevention play, at population scale. The addressable market for treating depression as a dementia risk factor is measurably larger than the market for treating depression alone. Read the WHO guideline

💡 Only 52% of evidence-based mental health apps are actually available to download

A systematic review of 81 self-guided apps with a rigorous evidence base found only 52% could be downloaded by a consumer today — mostly free, on both major app stores. Effect size and study quality had no significant association with commercial availability. The research-to-market pipeline in digital mental health is leaking more than half its output — a different version of the “proof isn’t payment” problem from issue #156. Read the review

Enjoy the summer. We will be back in September.

Are you working on the digital layer for the new interventional psychiatry — preparation, integration, monitoring, therapist training? Or building elsewhere in the mental health stack?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose, and with some sunlight,

The Masawa Team

🍄 Masawa Minute 157 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing: short reads, big questions, glimpses of where purpose meets performance.

Evidence is arriving in mental health at an unusual pace. Five studies and research programmes published recently span nearly 10,000 adolescents, more than 30,000 closed patient episodes, a 52,929-person employer-benefit cohort, and a six-country European feasibility programme. Together they show the kind of methodological seriousness the field has historically lacked, and we find that genuinely encouraging.

The harder question is what happens after the study is published. Clinical effectiveness and commercial viability are two different problems, and passing the first tells you nothing about the second. Digital mental health already has two sharp examples of companies with FDA clearances, real trial data, and genuine clinical need behind them that never found a payer willing to fund them at scale. For us, doing diligence on European mental health companies, the gap between what a study shows and what a statutory insurer will reimburse is the central question.

This Week’s Story

💳 Strong Evidence, but Who Will Pay for It?

Several outcome studies published recently push the field in a direction it needs to go.

• Grow Therapy’s retrospective cohort of nearly 10,000 adolescents showed 51.3% clinically meaningful improvement in depression and 57.8% in anxiety.
• Concert Health’s dataset, 30,162 closed patient episodes from 2018 to 2025 including 5,693 with severe depression, found 52% achieving clinically meaningful response and 21% reaching remission in the severe cohort.
• Ash’s single-arm observational study in 1,284 coaching users reported improvements in life satisfaction, sleep, and behavioral activation (no control group, not peer-reviewed).
• A Spring Health employer-benefit cohort of 52,929 adults across 589 US employers found 92.3% reliably improved on depression or anxiety, with effect sizes exceeding meta-analytic benchmarks for psychotherapy.
• And across Europe, the Horizon-funded ADVANCE project reported feasibility results from six sites, from a CBT chatbot in Lithuania to a climate-distress intervention in Germany, and is now moving into large-scale RCTs.

More rigorous evidence in mental health is welcome, and overdue.

A study proving a treatment works says nothing about whether the company behind it can build a viable and sustainable business from that proof of scientific evidence. In digital mental health, that gap already made companies vanish.

When FDA Clearance Wasn’t Enough

Pear Therapeutics built what the US regulatory system said was the right thing to build. reSET received FDA clearance in September 2017 for substance use disorder, the first prescription digital therapeutic in US history. reSET-O followed for opioid use disorder, then Somryst for chronic insomnia. Real randomized controlled trial evidence underpinned each product. Pear priced its treatments at over $1,000 per course and modeled reimbursement in the range of $5,000–$10,000 per treatment episode.

No payer created a billing code, and in April 2023, Pear filed for bankruptcy with solid evidence and no way to get reimbursed for it.

Akili Interactive followed a similar arc. EndeavorRx, the first video-game-based treatment for pediatric ADHD, received FDA authorization in 2020, backed by real trial data. By Q2 2023, Akili posted $114,000 in revenue against $15.3 million in expenses. The same reimbursement wall was held. In 2023, Akili cut roughly 40% of its workforce, abandoned the prescription model, and pivoted to an over-the-counter direct-to-consumer subscription called EndeavorOTC. The company was eventually sold to Virtual Therapeutics for approximately $34 million.

Both companies had FDA clearance and real trial data. What neither had was an answer to who would actually pay.

This is not only a US lesson. At Venture Café London’s early-stage validation panel this week, chaired by Dr Isabel Van De Keere, Masawa’s venture partner Dr. Youssof Oskrochi was asked to name the most dangerous metric in healthtech validation. His answer was the regulatory certificate presented as irrefutable proof: a Class III marking, he noted, did not stop a cardiac assist device that had been on the market for years from being shown unsafe. A certificate tells you a bar was cleared, not that a product behaves, or sells, as intended, which routes straight back to the harder questions it lets founders skip: who is actually affected, how big is the problem really, and is it painful enough that someone will pay to solve it? (see his LinkedIn post here)

Where the Architecture Is Working: Three European Pathways

The US experience matters precisely because Europe has built something structurally different. Three pathways now demonstrate that evidence can unlock payment when the regulatory architecture is designed to make that connection. The numbers behind them show why no single pathway should be treated as a finish line.

Germany’s DiGA fast-track: The Digitale Gesundheitsanwendungen framework, administered by BfArM, gives manufacturers a provisional listing for up to 12 months, extendable by a further 12, while they gather real-world evidence of a “positive care effect.” Meeting that bar converts the listing to permanent, and statutory health insurers pay €200–250 per patient, generally covering 90 days of use (source). Real mental health applications have made it through: Selfapy for depression, deprexis, MindDoc, and Velibra for anxiety all hold reimbursed status today (BfArM DiGA directory). But the gate is narrow. As of November 2025, of 239 DiGA applications submitted, only 57 received a positive assessment, 28 were assessed negatively, and 127 were withdrawn before a decision. Sixteen more were deleted from the directory after being listed. Roughly one in four applicants reaches a positive assessment.

France’s PECAN pathway: PECAN (Prise en charge anticipée numérique) routes therapeutic digital medical devices through CNEDiMTS for clinical benefit evaluation before attaching a fixed state tariff of €435, including tax, for the first three months, then €38.30 per month, capped at €780 per patient per year (source). That predictability is genuinely useful, and in July 2026, the first company completed all of the hurdles: Poppins, a game-based therapy for childhood dyslexia backed by an RCT (Poppins-02, involving 306 patients), became the first digital therapeutic approved through PECAN, with reimbursement expected to start this autumn. It is also strictly national: an approved DiGA in Germany does not carry over into PECAN in France. There is no European reimbursement passport. Each country runs its own review on its own timeline, so a company that clears Germany still faces every other country’s clearance from scratch.

UK NICE Medical Technology Guidance: Big Health’s Sleepio, digital CBT for insomnia, became the first digital therapeutic to receive NICE guidance, in 2022, built on 28 studies including 12 RCTs, with NICE confirming clinical and cost effectiveness and peer-reviewed trials showing 76% of patients achieving clinical improvement. That recommendation turned evidence into a commissioning signal NHS bodies could act on. NICE guidance does not guarantee adoption, though: NHS trusts can and do decline to commission technologies it has recommended.

These three pathways are real, and they’re the clearest evidence available that evidence-to-payment works in principle in European mental health. On their own, they don’t add up to a reliable, repeatable go-to-market: processes are slow, attrition is high even for strong evidence, and a win in one country does not travel to the next. Betting a company’s commercial future on a single national public pathway concentrates its regulatory and timeline risk into one outcome it can’t control.

What We Look For

Given Masawa’s own experience screening hundreds of companies in this space, the diligence lesson is straight-forward: we look for companies that treat reimbursement as one potential revenue stream among several, never the entire plan. The companies that earn our attention can also reach employers, private clinics, pharma and other health providers, or consumers directly, designed from the start rather than bolted on after a reimbursement bid stalls. Reimbursement wins, when they land, accelerate growth. They are not the prerequisite for survival.

Getting into one reimbursement scheme isn’t the holy grail. No matter how compelling its evidence, a company that cannot establish a clear path to monetization will fail to sustain its intended impact. Furthermore, relying entirely on a single revenue model merely delays this challenge rather than truly resolving it.

Masawa Spotlight

🎙️ Joshua on the Explorers Podcast

Joshua joined Lila Behr for Episode 5 of the Explorers Podcast, hosted by the We Are Impact Collective. Beyond the twenty-year path that led him to Masawa, the conversation focused on the innovative approach at the fund’s core, Nurture Capital, and why it’s key to delivering alpha. Listen to the episode

🏎️ Rally Pitch Comes to Jyväskylä

Masawa’s portfolio company Psyon Games, the Finnish applied-games studio behind FULL ADHD and other research-backed health awareness titles, is hosting a Rally Pitch event at Tommi Mäkinen’s HQ in Jyväskylä, Finland on July 29. The concept is literal: founders sit strapped into the passenger seat of a rally car, pitching their startup at maximum velocity while a professional driver maneuvers through the corners. Are you ready to pitch at full speed? Learn more

🤝 A Follow-Up: Human Change Foundation on Artificial Intimacy

Following our last issue on the regulatory wave targeting AI companions aimed at minors, we want to spotlight our friends at the Human Change Foundation and their Beware of Artificial Intimacy campaign. Their work (and their newsletters) make the case that AI built to simulate closeness poses a distinct risk to kids, and pushes for age limits, safety-by-design defaults, and a cultural shift back toward real human connection. Learn more

What Else We’re Reading…

🧒 Under-13s Are Mental Health’s Blind Spot

Imperial’s “New Childhood” report found children 13 and under are a stable, rising share of Shout’s crisis-text conversations, with most going on to return for support more than once — a population digital mental health tools are largely not built for. Read the full report

⚖️ EU AI Act’s High-Risk Deadline Slides to December 2027

The Digital Omnibus agreement pushes high-risk AI obligations back 16 months, giving clinical-AI builders breathing room, though the compliance work is only deferred, not reduced. Read the Council statement

🤖 Roughly 27% of AI Users Already Lean on AI for Mental Health Support

A new review confirms the number that’s about to show up in every pitch deck and regulatory filing this year — though the underlying studies range from 3% to 70%, a reminder that “AI for mental health” still means very different things to different researchers. Read the review

🧠 Connecticut Becomes the Fourth US State to Protect Neural Data

As of July 1, EEG readings from consumer wearables count as sensitive personal information under Connecticut law, with opt-in consent required and no threshold to trigger compliance. A likely bellwether for the rest of the BrainTech wearable category. Read the coverage

🏗️ Oxford’s £750M Mental Health Campus Clears Planning

A 950,000 sq ft mental health hospital, research centre, and Oxford graduate college at Warneford Park just got the green light — one of the largest single commitments of public and philanthropic capital to mental health infrastructure in Europe. Read the announcement

📋 How Europe Actually Pays for Digital Health

A Dutch Ministry of Health report maps financing and reimbursement models for digital health applications across eleven European countries — the most complete public survey we’ve seen on the exact mechanisms this issue’s main story is about. Read the full report

💰 Wellbees Raises $3.6M Series A to Expand in the Gulf

The AI-powered employee wellbeing platform — live in over 100 countries and 38 languages — closed a $3.6M Series A led by Qatar Development Bank to fund GCC expansion and a new Qatar office, a reminder that the employer channel keeps drawing capital while public reimbursement stays hard-won. Read the announcement

🤝 NeuroNation’s DiGA Becomes an Exit

Schwabe Group, the 160-year-old German pharma house behind Tebonin, took a majority stake in Berlin’s Synaptikon — whose NeuroNation MED app holds permanent DiGA reimbursement status — the clearest sign yet that clearing Germany’s reimbursement bar can end in a strategic exit rather than a slow bleed. Read the deal

In Closing

Are you building or backing mental health ventures and/or funds?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

💳 Masawa Minute 156 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, your bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

Over a ten-day period last month, a regulatory shift occurred across four jurisdictions at once: a social media ban for those under 16 in the UK, new mandatory crisis-intervention requirements for Canadian chatbots, a ban on AI-only therapy in Vermont, and a bipartisan US proposal to outlaw AI companion bots for minors. Far from being coincidental, these actions represent the initial legislative settlement of a fifteen-year debt incurred by unregulated attention capture, which is now finally being encoded into law.

⚖️ THIS WEEK’S STORY | The Attention Economy’s Mental Health Bill Comes Due

For fifteen years, the dominant model of digital interaction was designed for one outcome: to keep you engaged for as long as possible. The intentional design of algorithmic outrage, infinite scrolling, and dopamine-driven engagement was never a flaw; it was the fundamental architecture of the system. For over a decade, the resulting mental health toll on youth was clearly evidenced and widely discussed, yet it remained almost entirely unaddressed by legislative oversight.

That era is ending.

In the last two weeks of June, something shifted from policy aspiration to enforceable law across four jurisdictions simultaneously. The attention economy is receiving its first serious regulatory invoices. Now the question for investors in mental health is how to read it.

THE LEGISLATIVE WAVE: FOUR JURISDICTIONS, TEN DAYS

On June 15, UK Prime Minister Keir Starmer announced that children under 16 will be banned from major social media platforms — Snapchat, TikTok, YouTube, Instagram, Facebook, and X — with legislation expected before Parliament by the end of 2026. The announcement goes further than age verification: it simultaneously prohibits under-18s from accessing AI romantic companion chatbots, and bars strangers from contacting children on gaming and livestreaming platforms. The framing was explicitly targeted at children’s mental health measures.

Two days earlier, Canada introduced Bill C-34, the Safe Social Media Act, requiring chatbot services to be “safe by design” with mandatory crisis-intervention protocols for suicide and self-harm. The bill followed a devastating catalyst: a Canadian mother’s lawsuit alleging ChatGPT failed to respond when her daughter expressed suicidality 41 times across 18 months.

On June 17, Vermont Governor Phil Scott signed Act 156, making Vermont the second US state after Illinois to mandate a licensed mental health professional in the loop for AI-powered mental health services. Effective immediately: AI chatbots may not independently provide therapy, counseling, or therapeutic treatment.

And at the federal level, Representatives Foushee (D-NC) and Moore (R-UT) — with Senate companions Hawley (R-MO) and Blumenthal (D-CT), a notably bipartisan coalition — introduced the GUARD Act, which would ban AI companion chatbots for minors and create criminal penalties for violations. Seven additional state bills are simultaneously awaiting governor signatures, clustering around the same themes: protect minors, require human oversight, mandate self-harm protocols, require AI disclosure.

Far from a series of accidents, we are witnessing a deliberate, cross-jurisdictional alignment.

WHAT’S ACTUALLY BEING REGULATED — AND WHAT ISN’T

It is worth being precise about what the legislative wave is targeting. The distinction matters enormously for investors.

The legislation is overwhelmingly directed at two things: AI companion and entertainment chatbots deployed without clinical oversight, and social media platforms with addictive algorithmic design. It is not directed at clinical-grade AI tools operating within supervised healthcare settings. Vermont Act 156 explicitly exempts administrative AI — note-taking, scheduling, transcription. The AMA and APA, who both published “assistive not autonomous” standards within eight days of each other this month, are equally clear: the standard applies to AI that makes clinical decisions, not to AI that supports licensed clinicians who make those decisions.

The structural framework currently being adopted by both legislative bodies and clinical associations follows a distinct three-tier flow: AI facilitates the interaction, while a licensed professional maintains authority over review and ultimate decision-making. This three-layer model, live in products like Grow Therapy’s AI Coach and Talkspace’s newly launched Tee, is what the emerging regulatory framework protects — and what distinguishes it from the unguarded companion chatbots now facing legal exposure.

For mental health investors, this creates a structural moat. Companies that built clinical oversight into their architecture from day one are compliant by design. Those that launched without it face an expensive retrofit and a growing wall of regulatory risk.

THE LITIGATION FRONT

Legislation moves slowly. Litigation does not.

JCCP 5431, the coordinated chatbot litigation in California covering 13 cases against OpenAI, was assigned its trial coordination judge this month. The companion case Carrier v. OpenAI, filed June 11, alleges 41 unflagged suicidal disclosures before the plaintiff’s death. Discovery in these cases will establish evidentiary standards for chatbot safety that will effectively define the regulatory floor for the rest of the decade.

Separately, Spotlight PA testing found five AI companion chatbot platforms fabricating medical license numbers when probed. One vendor acknowledged this was intentional product design. The contrast with frontier models — which refused the same prompt — was stark. The litigation is not targeting AI as a category. It is targeting specific deployment choices made without safety infrastructure.

WHAT THE EUROPEAN TRAJECTORY LOOKS LIKE

The EU is not waiting for the US to settle. EU AI Act Article 50 chatbot disclosure obligations apply from August 2 — six weeks away. All EU providers of conversational AI must disclose their AI status to users; in sensitive contexts like mental health, the EU Commission’s Code of Practice (published June 10) suggests one-time disclosure is insufficient and periodic reminders may be required.

UK Technology Secretary Liz Kendall committed Ofcom to concrete chatbot regulations by the end of 2026, with the first restrictions potentially in force by spring 2027. The EU’s high-risk AI guidelines consultation — covering AI that causes “psychological trauma or serious mental health impacts” — closed July 10. EU member states must establish national AI regulatory sandboxes by August 2.

The direction is consistent across all five jurisdictions in question. The clinician-in-the-loop is becoming regulated infrastructure, not competitive differentiation. The question for every mental health AI company in Masawa’s pipeline is whether it was designed to meet this standard, or will need to be retrofitted to it.

WHERE MASAWA SITS

We have been arguing for several issues that the AI trust layer — companies building safe, supervised clinical AI infrastructure — is one of the defining investment themes of this decade in mental health. The legislative wave of the past two weeks converts that thesis from a forward-looking bet into a backward-looking description of what the market now demands.

This shift underscores our enthusiasm for founders anchoring their work in this domain, including Andy at RHISE.ai | Lucie, Jaime at CogniGuard, Rhagu and Larissa at Ginko, and Deepthi and Rupert at Neuly.

During our due diligence process, we employ a rigorous filter focused on three key questions, framed around integrity, ethics and representativity, based on the Responsible AI DD Tool from Project Liberty, ImpactVC, and Reframe Ventures. This assessment looks at who is interacting, who is conducting reviews, and who is the ultimate decision-maker? If a system relies solely on AI for any of these functions, if it misses the guardrails and escalation modes, and if ultimately gives decision power to the AI, it fails to meet the rising professional and regulatory benchmarks appearing across both the US and Europe. Organizations that have methodically constructed clinical-grade AI frameworks now find themselves in a position where their internal architecture has become the baseline legal standard, representing a true structural moat.

📍 MASAWA SPOTLIGHT

🇬🇧 THE REAL + Masawa Salon Dinner | London

We spent early June at THE REAL — the gathering of the global mental health investment and innovation community in London, and one of the events we most look forward to for the seriousness of its conversations.

Joshua moderated a panel on the Neuroscience of the Gut, exploring the emerging evidence linking metabolic health, gut biology, and neuroscience to mental health outcomes — and the investment implications of the gut-brain axis as a new therapeutic frontier.

Sabine moderated the Ethics of AI Companions and Digital Therapy panel — an increasingly urgent conversation given the regulatory wave already building at the time — and spoke on the funding models panel alongside Max Ahmed of the Wellcome Trust and Pooja Sikka of Meridian Health Ventures, examining how public, philanthropic, and market capital can be structured to support mental health innovation across the UK and Europe.

One moment worth noting: Julie Rubenstein of Better Society Capital opened the Mental Health VC Gold Rush panel by quoting the Masawa newsletter. We will take that.

The broader themes — trust as critical infrastructure, the case for resilience-first rather than disease-first research, the white space in women’s brain and reproductive mental health, the structural case for cross-border EU/UK clinical collaboration — felt unusually calibrated to where the market is actually moving.

We also hosted a cozy Masawa Salon dinner, an intentionally intimate gathering to sit with some of the harder questions. David Swanson of Prospira Global shared insights from their new Mental Health Integration for Impact report, which makes a clear case for treating mental health as core infrastructure rather than a peripheral theme. The conversation turned quickly from frameworks to responsibility as we discussed the topic of “as integration becomes the baseline, what future of mental health are we actually building toward, and what will it take for our own portfolios to reflect that?”

🇩🇪 SuperVenture | Berlin | Joshua

Joshua spent the following week at SuperVenture in Berlin, Europe’s annual gathering of VC fund managers and their LPs.

The structural fault line of European ventures came into sharp focus: EU pension funds allocate 0.01% of assets to VC, compared to 20% in the US — a capital supply gap that shapes everything downstream, from fund sizes to founder expectations to the pace of category creation. The Responsible AI session, the Women in VC Lunch, and the Future of AI and VC conversations reinforced a consistent theme: the next generation of defensible AI companies in Europe will be those with clinical rigour and regulatory clarity built in from founding, not those racing ahead of both.

🇫🇷 Les Premières Rencontres des Fondations Actionnaires | Paris | Sabine

Sabine attended the inaugural gathering of France’s “fondations actionnaires” community in Paris — an event that resonates directly with how we think about ownership, governance, and the long-term alignment of capital.

The “fondation actionnaire” model — in which a foundation holds a controlling or significant stake in a company, protecting it from short-term extraction while directing dividends toward philanthropic purpose — is well established in Northern Europe. Novo Nordisk’s foundation, the world’s largest enterprise foundation, owns both Novo Nordisk and Novonesis in perpetuity, managing €142B in assets and deploying 5% annually in grants. In France, Pierre Fabre, Schmidt Group, Groupe Monin, and Palais des Thés are among the pioneers of the model.

The question Sabine brought home: if foundation ownership goes mainstream, and the wealth created by companies is consistently redirected toward philanthropic ends rather than private gain, who funds the next generation of for-profit innovation? It is not rhetorical. Marie Ekeland’s 2050.do fund — a VC vehicle structured around steward ownership, where the management company is itself held in a perpetual fund with multi-stakeholder governance — is one answer. Masawa, with our innovative fund structure element, is another. The conversation is unresolved, which is exactly why it is interesting.

🇳🇱 HLTH Europe, Amsterdam — The One We Missed

We had real FOMO on HLTH Europe (June 15–18). From what reached us through the field: the AI team architecture conversation dominated the week — how Spring Health, Headspace, Grow Therapy, and Slingshot are structuring their clinical and engineering teams, and what “AI-native” actually means when you have payor relationships to protect. The consistent signal: companies where clinical and engineering leadership are co-located from founding operate fundamentally differently from those bolting AI onto existing clinical workflows. Next year, Amsterdam!

📚 WHAT ELSE WE’RE READING…

🛡️ Sword Health open-sources MindGuard — the clinical AI safety benchmark

Six weeks before EU Article 50 applies, Sword Health published MindGuard, its turn-level open-source safety classifier for mental health AI, including an honest account of its current limitations. It is positioning itself as the reference standard against which other clinical AI products will be measured. Read the technical post

💷 Wellcome commits £16bn to mental health science over a decade [UK]

CEO John-Arne Røttingen detailed Wellcome’s 2022–2032 strategy as the foundation nears its 90th anniversary: a £16bn commitment to mental health science over the decade. Positioning mental health as one of three core global priorities, the world’s second-largest health research funder is pivoting toward equitable distribution. By 2030, at least 25% of its active research portfolio will target Africa, South Asia, and Southeast Asia — a deliberate move to align scientific capacity with the regions where the mental health burden is most severe. Read the Wellcome strategy

🏛️ England opens consultation on a 10-year mental health strategy, prevention at its core [UK]

An eight-week evidence call (closes July 10) under Professor Peter Fonagy explicitly frames the shift from late diagnosis to earlier intervention and prevention. For investors: earlier intervention at population scale is the largest unaddressed market in UK mental healthcare. Read the announcement

🇫🇷 France’s €655M sovereign AI package includes a national clinical AI track [EU]

Prime Minister Lecornu’s announcement covers a French-language frontier model, public-sector deployment, and a dedicated clinical AI track including an AI assistant for the national insurer Ameli. France simultaneously ended its Palantir contract — a deliberate signal on sovereign AI infrastructure and data. Read the announcement

🧒 InStride Health closes $30M Series C for pediatric anxiety and OCD — a European white space [US]

InStride (virtual provider for children and teens with anxiety and OCD) raised a Series C led by Echo Health Ventures and General Catalyst, bringing total capital to ~$86M. No direct EU equivalent at this scale yet. Youth-focused, evidence-based, virtually delivered: this is a category waiting to be built in Europe. Read the announcement

📊 Mental health leads all US telehealth diagnoses — in every age group, every region [US]

FAIR Health’s Q1 2026 tracker: 52.1% of telehealth patients received a mental health diagnosis, the top category across all age groups and all US regions. Overall telehealth use rose 10.1% quarter-on-quarter. EU systems are directionally similar — therapist scarcity makes digital delivery structurally necessary, not optional. Read the FAIR Health report

🏆 Psychplus announces the acquisition of Koa Health to create world’s largest technology-enabled mental health company

PsychPlus, a leader in psychiatric and therapeutic care, has acquired Koa Health, a multinational digital mental health company, creating what it claims is the world’s largest technology-enabled mental health platform. The combined entity now serves over six million patients across the US, Europe, Australia and Asia-Pacific. Read the announcement

Are you building or backing mental health AI tools that put clinicians in the loop — or grappling with what the attention economy’s regulatory reckoning means for your portfolio?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better — invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

📵 Masawa Minute 155 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing: short reads, big questions, glimpses of where purpose meets performance.

This month, the Lancet put a number on the stakes: 1.2 billion people now live with a mental health condition, the world’s leading cause of disability. The easy read is a sickness story that needs a bigger fix.

We read it differently. With 91% of those people untreated, no amount of clinical capacity gets a population to well. The larger opportunity sits beyond the medical layer, in the technologies that help people flourish, and increasingly beneath the AI layer, in the tools that protect us from tech’s own harms. This issue makes the market case for both.

This Week’s Story

🌱 The Opposite of Illness Isn’t “Not Ill”

Fewer than one in ten people with a mental health condition receives adequate care. That figure, from the IHME’s 2023 Global Burden of Disease update, sets the scale. It also points to the wrong opportunity.

1.2 billion people now live with a mental health condition, up 95% since 1990. Mental ill-health has become the world’s leading cause of disability, ahead of cardiovascular disease and cancer. Major depression rose 131% over that period; anxiety 158%. With a 91% treatment gap, increased clinical infrastructure, however necessary, won’t solve this at population scale. You cannot treat your way to 1.2 billion well people.

That gap changes the investment logic. The opposite of illness isn’t “not ill.” It’s a person with energy, focus, connection, and agency. Prevention, resilience, cognitive performance, social connection: this is the flourishing layer. It sits beyond the clinical one, it touches everyone, and it has always been the larger market.

The 2026 Deals Make the Case Clear

The proof is now on the flourishing side.

• Oura, Finland’s sleep, recovery, stress, and metabolic-health wearable, filed confidentially for a US IPO with reported ambitions north of €9B. Its last private round (October 2025) valued it at $11B, up from $5B a year earlier, and it now projects ~$2B in 2026 revenue across 5M+ paid members. Days later it launched Ring 5 alongside a Brain Health Study with Cambridge Cognition that pairs in-app cognitive tasks with long-term biometrics, the first wearable mental-health dataset of its kind at scale. It’s the marquee European flourishing-market liquidity event of this cycle.
• WHOOP, the wearable-as-a-service platform for recovery, strain, and sleep, is raising ahead of a likely IPO; its WHOOP Physician Services arm was selected into the CMS Innovation Center ACCESS program, a reimbursed Medicare pathway launching July 2026. That is wellness technology entering reimbursed care for the first time.
• Headspace launched “Headspace for Cigna Healthcare” on 1 January 2026, free to 7M+ Cigna members, and PitchBook named it a 2026 IPO candidate. Calm now reaches ~39M covered lives across employer, health-plan, and provider networks.
• Nourish raised $100M Series C at a $1.75B post-money at the nutrition and metabolic-health intersection, squarely on the prevention side.
• Overall, the mental-wellness app market sits at ~$9.45B in 2026 and grows about 15% a year. Meditation apps alone are ~$2.4B.
• The clinical side is consolidating too. Universal Health Services is acquiring Talkspace for ~$835M, and Spring Health is buying Alma in a deal reportedly valuing the combined company at $6–7B.

The whole sector has reached liquidity. But the flourishing layer is the larger and faster-growing one.

Two Lanes, One Thesis

The “why now” runs in two directions at once.

Lane one is AI-enabling companies that expand mental health access and personalize support at scale: wearable-native platforms that trigger interventions from physiological biomarkers, screening tools that match clinician-level recommendation accuracy. The technology exists, and the outcomes are starting to land: Limbic’s clinician-supervised AI, live across NHS Talking Therapies, reported 42% more sessions attended and 25% higher recovery rates in a multi-site trial. The open question is whether it gets deployed in systems built to support users or to retain them: are we building tools for wellbeing and patient outcomes, or for engagement and retention metrics?

That question defines lane two: companies that dampen AI’s documented harm. The evidence is peer-reviewed now. Anthropic’s self-audit of ~1.5 million Claude conversations found roughly 1 in 1,300 distorted a user’s grip on reality through sycophancy. A Columbia study (Jutla et al.) found ChatGPT’s free tier gave inappropriate responses to psychotic-symptom probes 26 times more often than appropriate ones. UCSF psychiatrist Keith Sakata has reported treating a dozen patients with what’s now called “AI-associated psychosis,” a clinical entity recently set down in the first peer-reviewed case report. The World Health Organization now classifies generative-AI use as “a public mental-health concern”. Regulators are responding through the EU AI Act, EU Digital Services Act, Connecticut’s SB 5 AI-companion safeguards, and Colorado’s HB 1195.

A trust layer is forming to meet that harm. The Path’s custom model scored 95/100 on the Vera-Mental Health AI safety benchmark, against a ceiling of 65 for consumer chatbots. MindBench launched a community benchmark that scores frontier models against clinician and lived-experience consensus, with model cards tracking what each AI lab tests for in mental-health safety.

In an AI-saturated world, backing the technology that protects human flourishing from the advancements of technology itself is a full thesis on its own.

Where Masawa Sits

Mental illness is core to what we do; the burden numbers make any other position indefensible. What’s new is that we are naming and leaning into the adjacent layer: the flourishing opportunity beyond the clinical layer, and the trust infrastructure beneath the AI layer. We back the AI-enablers and the AI-harm-dampeners, and the best companies build for both at once.

The market is past the question of whether mental health is investable. Where to enter is what’s left, and our position hasn’t changed: we invest in technologies that enable all humans to thrive and flourish, supporting three main subject matters:

• Human potential: healthcare, fitness, longevity and healthy aging, nutrition, women’s health, human performance
• Meaningful connection: community, education and cognitive security, financial wellness, employee, family, parenting, friendship and couple wellbeing
• Regenerative living systems: democracy-tech, social cohesion, climate resilience, built environment, urban space design

Masawa Spotlight

🌍 Three Days at Katapult Future Fest

We spent last week at Katapult Future Fest in Amsterdam: investor day on the waterfront at Tolhuistuin, then the festival transmuted to Ruigoord, the free-spirited artists’ village on the outskirts of the city, with conversations spilling into the gardens between sessions.

It leans unapologetically “woo-woo,” and that’s exactly why we keep coming back. People ask the uncomfortable, political, and philosophical questions that finance usually skirts.

A few stuck with us. “What LPs and GPs Won’t Say Out Loud” was an honest look at the human realities beneath capital allocation: trust, ghosting, status, belonging, and our complicated relationship with money. In “Economic Circus: How Rich Is Too Much?”, Kate Raworth asked whether a society that draws a poverty line should also draw an extreme-wealth line; the room voted, and landed on €10 million. And an Oxford-style debate, “Is Defence an Investment in Humanity?”, weighed whether security is a precondition for flourishing, and whether impact capital can back defense without becoming dependent on conflict for returns. No easy answers, just unusually nuanced arguments. The “Esoteric Investing” session gave a few new non-mainstream due diligence tools as we assess our investment opportunities.

We left more convinced than ever that the flourishing layer and the AI-trust layer are the next decade’s thematics where impact and returns will be made.

Read Sabine’s insights on her LinkedIn post.

What Else We’re Reading…

🏛️ England launches a once-in-a-generation mental health strategy, with prevention at its core

The UK opened an eight-week call for evidence (closes 10 July) for a 10-year plan, with an independent review under Prof. Peter Fonagy aimed at shifting the system from late diagnosis toward earlier intervention. Read the full announcement

⚖️ EU AI Act: high-risk health-AI compliance pushed to December 2027

The Digital Omnibus deal gives health-AI builders 16 extra months of runway from August 2026, while adding new supply-chain obligations. A clear signal for anyone deploying AI in clinical workflows. Read the Council statement

🔬 Digital CBT beats referral in a 26-college trial, and the effect held at two years

A population trial (39,194 students screened) found that screening plus guided digital CBT cut any-disorder prevalence at two years and more than doubled care uptake, 74% versus 30%, in particular among minority groups. Evidence that scalable, preventive tools work. Read in Nature Human Behaviour

🏆 Wellcome and Nature launch the world’s largest prize for mental health science

$1M to the winner, $250K to three finalists, for rigorous, lived-experience-informed advances in anxiety, depression, and psychosis. Applications close 18 September. A structural tailwind for the research-to-venture pipeline. Read the announcement

📈 Behavioral health in 2026: the pivot from growth to proof

The sector’s coming-of-age. Capital is back, but clinical evidence and outcomes are now the competitive edge, replacing user-count growth. Read the full article

📱 How people actually use mental health apps

Real-world usage data on how engagement with mental health apps holds up outside the lab, a useful reality check for the engagement-versus-outcomes question at the heart of our lane one (see above). Read the study

🛡️ Independent safety ratings for AI mental health apps

Common Sense Media now rates AI mental health apps for safety and trustworthiness, exactly the kind of consumer-facing trust signal the new AI-safety layer is built to provide. See the ratings

🧬 Pathways to longevity

Harvard Health on the science and strategies behind healthy aging, the flourishing layer applied to a longer, healthier life. Read the article

In Closing

Are you building or backing ventures that help people flourish, not only recover?

Curious how we approach blended value, systemic change, resilience, and impact measurement?

💌 Let’s talk. Or better, invest in Masawa or co-host a Masawa Salon with us in your city or with your community.

With purpose,

The Masawa Team

🌱 Masawa Minute 154 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.

Welcome back to the Masawa Minute, our bi-weekly pulse on mental health, social impact, and impact investing — short reads, big questions, and glimpses of where purpose meets performance.

This week, one report stood out in particular. The OECD released The Economic Case for Preventing Mental Ill-Health — a landmark publication quantifying something society has underestimated for far too long:

• mental health is not simply a healthcare issue,
• it is an economic infrastructure issue.

And the cost of inaction is becoming impossible to ignore.

At the same time, conversations we joined across Europe — from the CEE Impact Forum in Munich to the first European Brain Economy Summit in Brussels — reinforced a broader shift already underway: mental health, brain health, inclusion, productivity, resilience, and economic competitiveness are no longer separate conversations.

They are becoming one.

🧠 This Week’s Story: The Economic Case for Preventing Mental Ill-Health

The OECD’s latest report may become one of the most important policy signals for the future of mental health investing. For years, mental health has often been framed as a social challenge, a healthcare burden, or a workplace wellbeing issue.

The OECD reframes it as a macroeconomic imperative.

According to the report, mental ill health affects more than one in five people across OECD and EU countries, resulting in one of the most significant public health and economic challenges. Among diagnosed cases, anxiety disorders are the most common, accounting for approximately 40% of all cases. These are followed by depressive disorders at 20% and substance use disorders — which include alcohol use disorders — at 17%.

The increase in mental health conditions has been particularly evident among adolescents and young adults, with more than one in four individuals aged 15‑24 currently experiencing a mental disorder. This is a cause for concern because when such conditions emerge early in life and remain untreated they often persist into adulthood, with lasting consequences for health and productivity.

OECD micro-simulation work estimates that major depressive disorders, generalised anxiety disorders and alcohol use disorders will lead to a 2.5‑year reduction in healthy life expectancy across the EU during 2025–2050. At the population level, this equates to approximately 28,000 premature deaths every year.

The financial implications are equally stark: mental health conditions are estimated to account for roughly EUR 76 billion in annual healthcare costs across the EU, which represents approximately 6% of total health budgets, while mental ill-health costs OECD economies the equivalent of 4–5% of GDP annually, driven by: reduced labour participation, absenteeism and presenteeism, chronic disease burden, educational underperformance, and long-term productivity loss.

Despite multiple policy frameworks across countries, a massive treatment gap remains. Nearly two‑thirds of individuals in the EU who require mental healthcare are estimated to face inadequate access to services. This stands in sharp contrast to the 3.8% of people who report unmet needs for general medical treatments.

In other words, mental wellbeing is no longer peripheral to economic performance. It is becoming foundational to it.

🌍 Prevention Is Finally Entering the Economic Mainstream

One of the report’s strongest conclusions is that prevention generates measurable economic returns, and importantly that it is no longer viewed solely through a public-health lens.

It’s investable, measurable, and scalable.

Scaling up evidence‑based interventions offers a cost-effective pathway to better health outcomes. The OECD highlights areas such as workplace mental health, youth prevention, digital interventions, community-based support, social prescribing, and early behavioural care. For example, cognitive behavioural therapy remains a cornerstone of treatment and prevention, helping individuals reframe negative thinking patterns. Other successful approaches include mental health literacy programmes, mindfulness-based techniques and digital tools that provide remote support for mild to moderate symptoms.

This mirrors what we observe across the venture ecosystem: the market is shifting from reactive care toward proactive resilience infrastructure, from crisis management to prevention and cognitive performance, and from fragmented wellness tools to integrated human resilience systems.

And capital is beginning to follow.

🧠 The Rise of the Brain Economy

Interestingly, this broader evolution was echoed repeatedly at the first European Brain Economy Summit in Brussels, where Joshua joined researchers, policymakers, investors, and health leaders discussing one core question:

“What will it take to turn Europe into the world’s leading Brain Economy?”

One message became clear: Europe does not lack talent, science, or institutions. It suffers from an “execution gap.”

Discussions repeatedly returned to: ageing populations, declining cognitive resilience, rising mental health burdens, neurotechnology, youth brain capital, and the need to better connect public and private capital.

Several speakers also emphasized that Europe still dramatically underinvests in brain and mental health relative to other sectors like oncology.

And perhaps most importantly:

“We must not differentiate between mental health and brain health.”

This strongly resonates with what we discussed in Newsletter #151 around Brain Capital:
cognitive resilience, emotional wellbeing, creativity, adaptability, and social cohesion are becoming core economic assets in the age of AI.

💡 What This Means for Investors

At Masawa, we know the next decade will not simply produce “mental health startups.” It will produce companies building:

• human resilience infrastructure,
• cognitive infrastructure,
• behavioural infrastructure,
• and brain-positive systems.

And the investment opportunity is expanding beyond healthcare, because mental wellbeing now shapes workforce productivity, educational outcomes, leadership quality, innovation capacity, social cohesion, and democratic resilience.

The OECD report simply gives economic language — and political legitimacy — to what many founders and operators already intuitively know.

Healthy societies are built by healthy minds.

🔦 Masawa Spotlight

🌍 CEE Impact Forum: Driving Change, Delivering Returns

On Wednesday, Sabine joined the CEE Impact Forum in Munich to discuss the evolution of impact investing across Central & Eastern Europe, a region emerging as a meaningful innovation and impact corridor.

One key takeaway from the discussions — and from the newly released CEE Impact Report 2026 by KPMG — is that the narrative around impact investing in CEE is finally becoming data-backed. The report highlights:

• 61 impact enterprises across Austria, Hungary, Croatia, and the Czech Republic,
• strong revenue growth,
• improving capital efficiency,
• and growing LP appetite for impact strategies delivering market-rate returns.

Most striking: sampled impact enterprises reported a 45.5% median 3-year CAGR and generated €1.57 of revenue per €1 of capital raised.

Another insight that strongly resonated with us:

“Too much narrative, too little data.”

As impact investing matures, evidence, accountability, and measurable outcomes are becoming prerequisites and not optional extras.

And Health & Wellbeing already represents one of the largest impact categories in the region.

🧩 Mental Health Integration for Impact

This week, we were also proud to see the release of Mental Health Integration for Impact, a new report by Prospira Global — founded by Masawa Venture Partner Elisha London.

The report reinforces a conviction we have held since day one at Masawa: mental health is not a standalone category, but foundational infrastructure for women’s health, child development, climate resilience, productivity, and social cohesion. As highlighted in the report, mental health is “a cross-cutting determinant of whether funding strategies achieve their intended impact.”

This report examines these three domains as practical entry points for integration, unpacking how mental health interacts with women’s health systems, child and youth platforms, and climate financing in distinct but overlapping ways. It finds that while interest in integration is widespread, in practice this is often informal in nature and unfunded.

It is encouraging to see global institutions increasingly move from awareness to implementation, embedding mental health directly into funding architectures and resilience strategies across sectors. A strong validation of the direction in which the ecosystem is evolving.

🤝 Let’s Meet (or Join Us)

We will be on the ground across Europe in the coming weeks and would be delighted to connect in person or host you at one of these gatherings:

• Amsterdam (May 26–29) — Katapult Future Fest (Joshua & Sabine)
• London (June 3) — The REAL Summit (Joshua, Sabine & Youssof)
• Berlin (June 8–10) — SuperVenture (Joshua)
• Berlin (June 9) — World Fund Summit (Joshua)
• Paris (June 9) — Premières Rencontres des Fondations Actionnaires (Sabine)

We have a limited number of tickets available for some of these events. If you’d like to claim a ticket or schedule a 1:1, reply by Friday May 22nd, so we can prioritize your slot.

⏳ Masawa Fund I: Fundraising Close

The opportunity to invest in Masawa Fund I ends on June 30th, 2026.

📚 What else we’re reading

🤖 Limbic crosses 700,000 NHS patients supported
London-based Limbic continues setting the benchmark for clinically validated AI mental health tools in Europe, now deployed across 63% of NHS Talking Therapies services and 13 US states. A strong signal that evidence-based AI mental health infrastructure is moving into the mainstream. Read here

🧠 One Mind Accelerator 2026 Cohort: Companies to Watch
The 2026 One Mind Accelerator cohort brings together 14 early-stage mental health startups across neurostimulation, social prescribing, family care, addiction, AI diagnostics, and evidence generation. We’re especially watching Flow Neuroscience, SocialRx, Manatee Health, Vitalic Health, Birches Health, Engram, and Radicle Science: strong US analogues to several themes we believe Europe should build around: clinical evidence, prevention, caregiver support, and scalable non-pharmacological care. Read here

🕶️ brainjo Raises €2M for VR-Based ADHD Therapy
Regensburg-based brainjo raised €2M led by High-Tech Gründerfonds (HTGF) to develop VR-based digital therapeutics for children with ADHD, with the goal of becoming a reimbursable DiGA in Germany. Read here

💫 Wisdom Ventures Closes $77.7M Fund II for AI-Native Wellbeing
Wisdom Ventures closed a $77.7M Fund II to back startups that leverage AI to improve health, wellness, and human connection — with high-profile backers including Reid Hoffman and Dr. Vivek Murthy joining as Senior Venture Partner. This strongly validates the “wellbeing infrastructure” framing Masawa has been building in Europe. Read here

🇳🇱 OpenUp Raises €20M to Scale Employee Mental Health Across Europe
Amsterdam-based OpenUp raised €20M from Smartfin and Rubio Impact Ventures to expand its employee mental health platform across the Netherlands, Germany, the UK, France, and Belgium. With 2,000+ corporate customers, OpenUp’s round is one of the clearest signals that European workplace mental health is entering a consolidation and scale phase. Read here

🇦🇹 nyra health Raises €20M / $23M for AI-Driven Neurotherapy
Vienna-based nyra health raised a major Series A to scale its AI-enabled neurotherapy platform across Europe and the US. The company’s reimbursed digital rehabilitation model and integration into German standard care make it a strong example of where mental health, brain health, and digital therapeutics are converging. Read here

🇪🇺 European Mental Health Week 2026: “Stronger Together”
This year’s European Mental Health Week focused on “Stronger Together: Prioritise Mental Health in a Changing Europe.” EU-funded projects such as Joint Action MENTOR and Horizon Europe initiatives were highlighted as examples of cross-border mental health promotion and prevention. Read here and discover the EU Projects here

🤖 Stanford: AI Chatbots and “Delusional Spirals”
Stanford researchers analysed 19 human-chatbot transcripts and found how AI relationships can spiral when chatbots validate distorted beliefs without meaningful challenge or intervention. This is becoming essential reading for anyone underwriting conversational AI in mental health: defensibility is not a disclaimer, it is clinical architecture, escalation, and governance. Read here

🌡️ Wellcome: Heat & Mental Health Funding Call
Wellcome is funding research into how heat affects anxiety, depression, and psychosis, with a focus on the communities most impacted by climate change. This is directly relevant to Masawa’s climate resilience lens, especially for startups addressing eco-anxiety, heat stress, displaced populations, and youth wellbeing. Read here

⚖️ EU AI Act: Transparency Rules Coming Into Force
The EU AI Act’s transparency obligations for certain AI systems enter into force on 2 August 2026, including requirements for users to be informed when they are interacting with AI. Separately, Article 5 already prohibits certain harmful AI practices, including manipulation and exploitation of vulnerabilities. For any conversational AI mental health company, EU AI Act readiness should now be a standard DD item. Read here and for further information on the implementation timeline here

Closing Thoughts

Across policy circles, neuroscience, venture capital, and philanthropy, one idea seems to converge: the future economy will not only be shaped by technological capital. It will be shaped by human capital, aka: brain capital. relational capital. mental resilience.

And the societies that thrive tomorrow will simply be the ones that learn to protect — and invest in — the minds of their people today.

Wanna know how? 💌 Let’s talk. Or better: invest in Masawa, or co-host a Masawa Salon with us in your city.

With purpose,

The Masawa Team

🧠 Masawa Minute 153 was originally published in Masawa on Medium, where people are continuing the conversation by highlighting and responding to this story.