<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Propagator Ventures on Medium]]></title>
        <description><![CDATA[Stories by Propagator Ventures on Medium]]></description>
        <link>https://medium.com/@propagatorvc?source=rss-ab3eeb5caf3d------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*JB08gUQQZGyg6ZBXc8ZOBQ.png</url>
            <title>Stories by Propagator Ventures on Medium</title>
            <link>https://medium.com/@propagatorvc?source=rss-ab3eeb5caf3d------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 13:11:29 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@propagatorvc/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Better Medicine, Fast — Why We Invested in Intrepid Labs]]></title>
            <link>https://propagatorvc.medium.com/better-medicines-fast-why-we-invested-in-intrepid-labs-ea24b8a1cd9c?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/ea24b8a1cd9c</guid>
            <category><![CDATA[pharmaceutical-sciences]]></category>
            <category><![CDATA[drug-formulation]]></category>
            <category><![CDATA[machine-learning-ai]]></category>
            <category><![CDATA[automated-lab-systems]]></category>
            <category><![CDATA[drug-discovery]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Thu, 23 May 2024 20:26:59 GMT</pubDate>
            <atom:updated>2024-05-23T22:07:39.325Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IqXfe9cfo8AFbVQmiJTFlA.png" /></figure><h3><strong>Better Medicine, Fast — Why We Invested in Intrepid Labs</strong></h3><p><strong>TL;DR: </strong>Propagator Ventures has invested in <a href="https://intrepidlabs.tech/">Intrepid Labs,</a> a Canadian startup that combines machine learning, robotics, and pharmaceutical science to speed up drug development. Founded in 2023 and spun out of the <a href="https://www.utoronto.ca/">University of Toronto</a> (U of T), Intrepid Labs addresses the bottlenecks in drug formulation, which is often slow, costly, and labour-intensive. The cross-disciplinary team at Intrepid Labs includes deep domain expertise in AI, robotics, and drug formulation science. Intrepid’s technology can reduce drug development timelines, unlock new therapies, enable personalised medicine, and lower clinical failures. Propagator Ventures sees Intrepid Labs’ self-driving labs as the future of drug development, creating life-saving therapies while cutting costs and attrition rates.</p><h3><strong>Integrating machine learning, robotics, and pharmaceutical science to fast track the future of drug development.</strong></h3><p>Our recent investment in <a href="https://intrepidlabs.tech/">Intrepid Labs</a> marks our first foray into the emerging field of drug formulation driven by machine learning and aligns with our strategy to back transformative deep-technologies and world-class teams tackling grand global challenges. We are excited to partner with Intrepid’s <a href="https://intrepidlabs.tech/home/#team">visionary founders</a> and join <a href="https://radical.vc/">Radical Ventures </a>as ambitious first backers of Intrepid Labs in Canada.</p><p><strong>Rooted in Pioneering Research</strong></p><p>Intrepid Labs spun out of U of T in 2023 to commercialise ground-breaking medical innovation stemming from the labs of leading researchers in machine learning, robotics, and drug formulation science. Intrepid is the first company to emanate from the <a href="https://acceleration.utoronto.ca/">Acceleration Consortium</a> (AC), an institutional strategic initiative based at U of T, set up to accelerate the design and discovery of new functional materials using self-driving laboratories powered by AI. The Consortium is led by Intrepid Labs co-founder, <a href="https://acceleration.utoronto.ca/researcher/alan-aspuru-guzik">Prof. Alán Aspuru-Guzik.</a></p><p><strong>The Formulation Bottleneck</strong></p><p>While the pharmaceutical industry has seen incredible advances in drug discovery over the past few decades, a critical bottleneck remains — converting promising new therapeutic compounds into viable medical products that can be safely administered to patients. After a new drug candidate is identified through the discovery process, a formulation needs to be developed for further pre-clinical and clinical evaluation.</p><p>Formulation development is typically a time-consuming and labour-intensive process that is largely reliant on benchmarking against existing formulations and trial and error experimentation. As a result, companies often settle on a formulation that is “good enough”. But with over 90% of drug candidates failing due to efficacy or toxicity issues, this conventional approach limits our ability to unlock the full therapeutic potential of new drug molecules.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/448/1*fMWb4YRby-p68IL6XXkAgA.png" /><figcaption>There is an untapped opportunity for pharma companies to move better formulations into the clinic. Currently, 99.9% of final formulations are never explored before the release of a drug (<em>source:</em> Intrepid Labs).</figcaption></figure><p><strong>Intrepid’s AI-Driven Formulation Platform</strong></p><p>Intrepid Labs is pioneering a radical solution — a self-driving laboratory that combines AI, robotics, and high-throughput experimentation to autonomously optimise drug formulations. Their proprietary platform — called VALIANT— takes a data-driven approach that employs advanced machine learning and robotic automation to efficiently explore the large formulation design space and identify the optimal formulation for a specific drug. Intrepid’s data-driven approach aims to turn this high-dimensional challenge into a rapid optimisation problem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/406/1*rrcMwxHHPk86IL3dyrQAjw.png" /><figcaption>VALIANT is a modular, autonomous laboratory that leverages data-driven experiment planning to optimise formulation compositions and uses robotic laboratory equipment to prepare and characterise formulation candidates in parallel (source: Intrepid Labs).</figcaption></figure><p><strong>A Platform for Intelligent Drug Design</strong></p><p>VALIANT rapidly prepares and tests a wide array of candidates in parallel, using the results to iteratively refine its machine learning models and zero in on optimal compositions. This closed-loop system explores the design space orders of magnitude faster than conventional methods, using much less resources, identifying viable formulations for even highly challenging drug molecules.</p><p>The “build-test-learn” cycle, powered by proprietary machine learning models and wet lab automation, enables Intrepid to search a significantly larger formulation space and surface non-obvious solutions that may elude other drug developers. Crucially, VALIANT is modular and can optimise formulations based on various drug delivery technologies, and moreover it has the capability to discover entirely new drug delivery technologies. The platform has already demonstrated the ability to rescue failed compounds and rapidly optimise long acting injectables, oral solid dosage forms, and nanoparticle formulations for small molecules and biologics. This flexibility across therapeutic areas and product types provides multiple shots on goal.</p><p>Intrepid Labs’ deep scientific foundation, augmented by a robust IP portfolio, provides the company with an unparalleled technological edge in tackling the formulation bottleneck that constrains the pace and economics of drug development.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/488/1*DcpLClaUVSkaGSwLXkc2Wg.png" /><figcaption>VALIANT’s formulation preparation and characterisation workflows can be configured to support development of any formulation type, from injectable depots to solid dosage forms to nanoparticles for parenteral administration of drugs. Intrepid’s closed-loop experimentation approach requires &lt; 1 gram of drug sample to converge on lead formulation candidates, with throughput ranging from tens to hundreds of unique formulations per day (<em>source:</em> Intrepid Labs).</figcaption></figure><p><strong>Team and Ecosystem</strong></p><p>Beyond its computational capabilities, Intrepid’s edge lies in its <a href="https://intrepidlabs.tech/home/#team">cross-disciplinary team</a> with deep domain expertise at the intersection of AI, robotics, drug formulation &amp; development. <a href="https://allen.pharmacy.utoronto.ca/dr-christine-allen/">Prof. Christine Allen</a> is a world-renowned entrepreneur and merited drug formulation and development scientist, with more than 170 publications. Her co-founders include Prof. Alán Aspuru-Guzik, a serial company founder and internationally recognised pioneer at the interface of quantum information, machine learning and chemistry. The founding team also includes <a href="https://www.researchgate.net/profile/Pauric-Bannigan">Dr. Pauric Bannigan,</a> an accomplished drug formulation scientist and inventor with extensive experience in the development of innovative drug delivery platforms, and <a href="https://www.researchgate.net/profile/Riley-Hickman">Dr. Riley Hickman</a>, an expert in materials discovery, applied mathematics, machine learning and robotics.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kWpE2NVA7zxHb8GWcY1UyA.png" /><figcaption>Intrepid Labs’ stellar founding team (source: Intrepid Labs)</figcaption></figure><p>With decades of experience and domain expertise, connections across industry and stellar academic and commercial pedigrees this team has the perfect credentials to bridge the gap between bits and molecules and to realise the immense potential of their self-driving lab technology. As a resident of the <a href="https://jnjinnovation.com/locations/jlabs/jlabs-toronto#:~:text=Situated%20right%20in%20downtown%20Toronto,the%20advancement%20of%20healthcare%20innovation.">Johnson &amp; Johnson Innovation — JLABS</a> incubator in Toronto and member of the <a href="https://vectorinstitute.ai/programs/ai-startup-and-scale-up-program/">Vector Institute’s FastLane Program</a> Intrepid Labs is also set to benefit from its robust position in the vibrant Canadian tech ecosystem.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*AC2SliNsBfpxRoPGUqKTGg.png" /><figcaption>Intrepid Labs partnerships (<em>source:</em> Intrepid Labs)</figcaption></figure><p><strong>Accelerating Medicines that Matter</strong></p><p>The ability to optimise drug products with unprecedented speed and scale opens up significant opportunities for Intrepid across the $1.4T pharma value chain. When successful, their platform could:</p><p>● Dramatically <em>shrink drug development timelines</em> by rapidly optimising formulations for solubility, stability, and bioavailability.</p><p>● <em>Unlock new therapeutic modalities</em> by rescuing promising drug candidates previously abandoned due to poor pharmacokinetics.</p><p>● Facilitate <em>personalised and precision medicine</em> by tailoring formulations to specific patient populations.</p><p>● <em>Reduce clinical failures </em>by prioritising formulation optimization earlier in development.</p><p>● <em>Extend patent life and create new revenue streams</em> by reformulating existing drugs.</p><p>● <em>Reduce complexity and cost of </em>existing drug product development and manufacturing while maintaining performance.</p><p>The global pharmaceutical industry is steadily growing; even capturing a fraction by effectively streamlining the formulation dimension could be transformative for Intrepid Labs. Not only is their technology highly scalable — it can begin formulation optimization without any initial data points and learn through experience. This capability allows for the inclusion of the newest classes of therapeutic agents and eliminates the need for large datasets to deploy ML models. In a world with growing data security concerns, this provides a powerful competitive advantage.</p><p><strong>The Opportunity Ahead</strong></p><p>By making drug formulation more of a data-driven engineering problem than an artisanal process, Intrepid has the potential to reshape how medicines are designed, tested and delivered. While early, Intrepid is pioneering an emerging paradigm — applying machine learning, robotics, and autonomous systems to accelerate not only drug discovery, but the entire drug development pipeline.</p><p>Intrepid Labs’ initial formulation focus tackles a key bottleneck, but their vision is grander: The potential for reducing the need for time consuming and costly experimental studies by applying machine learning models to accelerate the design of long-acting injectables was described in<a href="https://www.nature.com/articles/s41467-022-35343-w"> a study published in Nature Communications last year</a>; one of the first to apply machine learning techniques to the design of polymeric long-acting injectable drug formulations.</p><p>We believe Intrepid’s self-driving labs represent the future of drug development, much like how AI is accelerating design cycles in materials, semiconductors and other complex product categories.</p><p>At its core, Intrepid is tackling one of healthcare’s most valuable challenges — accelerating the development of life-saving therapies while reducing costs and attrition rates. We at Propagator Ventures are excited to support Intrepid in ushering in a new era of programmable, intelligent drug design — where the pace of medicine is no longer limited by trial and error but accelerated by algorithms and automation. The opportunity to reduce the time and cost associated with formulation development by a factor of 10 is a game-changer not just for the industry but for patients everywhere.</p><p><em>Disclaimer:</em> This article has been edited by an AI model, but the content is original and was generated by human authors.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ea24b8a1cd9c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why We Invested in Koala Tech Inc.]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-koala-tech-inc-62b482c46809?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/62b482c46809</guid>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[deeptech]]></category>
            <category><![CDATA[japan-startup]]></category>
            <category><![CDATA[oled]]></category>
            <category><![CDATA[optoelectronic-devices]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Tue, 24 Oct 2023 10:21:14 GMT</pubDate>
            <atom:updated>2024-09-05T11:35:14.495Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>Why We Invested in Koala Tech Inc.</strong></p><p><strong>Beyond OLED — Organic Semiconductor Laser Diode (OSLD) technologies for next-generation displays, optical sensing, biometrics and more.</strong></p><p>Two years ago, we were introduced to <a href="https://www.linkedin.com/in/chihaya-adachi-20493256/">Prof. Chihaya Adachi,</a> <a href="https://www.linkedin.com/in/fatima-bencheikh-a26143117/">Dr. Fatima Bencheikh</a>, <a href="https://www.linkedin.com/in/takao-hayama-4865b931/">Takao Hayama</a>, and their colleagues who had recently founded <a href="https://koalatech.co.jp/">Koala Tech Inc</a>., short for Kyushu OrgAnic Laser Technology. Their aim was to further develop and commercialize a long-elusive kind of optoelectronic device based on organic semiconductor laser diode (OSLD) technology. <a href="https://iopscience.iop.org/article/10.7567/1882-0786/ab1b90">This technology had recently had its practical viability demonstrated</a> — as a world first — by Prof. Adachi’s research group at <a href="https://www.kyushu-u.ac.jp/en/">Kyushu University</a> in Japan. As an innovative platform technology organic semiconductor laser diodes will have a transformative impact in telecommunications, medical diagnostics, and display technologies, with potential for far surpassing the capabilities of existing organic light-emitting diodes (OLED) technology.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bEj3-U49j1_GobZ4xXc7XQ.png" /><figcaption>Center for Organic Photonics and Electronics Research (OPERA), Kyushu University.</figcaption></figure><p>In our many conversations, the Koala Tech team — later bolstered by the addition of CFO <a href="https://www.linkedin.com/in/shinobuito/">Shinobu Ito</a>, R&amp;D Manager <a href="https://www.linkedin.com/in/yoko-fukunaga-5a911658/">Yoko Fukunaga</a> and R&amp;D Leader <a href="https://www.linkedin.com/in/morgan-auffray-931771b8/">Morgan Auffray</a> — built a convincing case for the promise of OSLDs in a wide range of applications. With their strong credentials and promising results, the team is probably the best positioned in the world to bring OSLD technology from the lab to the market. OSLDs represent a foundational technology set to expand the application of lasers in fields ranging from displays and healthcare to biometrics, augmented and/or virtual reality, and much more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Bfh1ma-ywuO4AtFIxjRkHw.jpeg" /><figcaption>Professor Chihaya Adachi, Director of the Center for Organic Photonics and Electronics Research (OPERA) at Kyushu University. Photo credit: Koala Tech Inc.</figcaption></figure><p>Our recent investment in Koala Tech marks our first enterprise outside of Europe or North America and aligns with Propagator Ventures’ strategy to support the foremost technologies and teams in the world. Notably, we became Koala Tech’s first investor from outside Japan. We have joined existing notable Japanese investors, including Sony Group Corporation and Beyond Next Ventures (BNV), in a financing round concluded this summer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Ynt0sz4tuhN4DAlbDGmSPQ.jpeg" /><figcaption>Koala Tech CEO Dr. Fatima Bencheikh (right) and CFO Shinobu Ito (left). Photo credit: Koala Tech Inc.</figcaption></figure><p><strong>Rooted in Pioneering Efforts</strong></p><p>This investment is rooted in the pioneering efforts of the <a href="http://www.cstf.kyushu-u.ac.jp/~adachilab/lab/?page_id=1442">Adachi and Nakanotani Laboratory</a> at the <a href="https://hyoka.ofc.kyushu-u.ac.jp/search/organization/11590000/english.html">Center for Organic Photonics and Electronics Research (OPERA)</a> at Kyushu University in Japan, a global leader in the field of organic electronics. <a href="https://en.wikipedia.org/wiki/Chihaya_Adachi">Prof. Chihaya Adachi</a> is arguably one of the world’s foremost authorities on organic electronics devices and a leading innovator and pioneer whose work is considered as fundamental to the advancement of OLED technologies.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F9ZpOpAE7dcRZoD_fauikw.jpeg" /><figcaption>Center for Organic Photonics and Electronics Research (OPERA), Kyushu University.</figcaption></figure><p><strong>Understanding OSLD: A Post-OLED Technology?</strong></p><p>Organic laser diodes, which use carbon-based organic materials to emit laser light, pave a promising pathway to applications unattainable by their inorganic counterparts used in traditional devices. OSLDs hold potential for being more cost-effective and adaptable compared to existing inorganic laser diodes. They also facilitate the development of novel optoelectronic devices that surpass many current OLED-based solutions.</p><p><strong>Advantages and Applications of OSLD</strong></p><p>OSLD can hold promise in developing high-resolution display technologies and advanced medical devices, potentially revolutionizing diagnostic and treatment methodologies, biometrics, smart displays, optical sensing, and more. The technology can also open new avenues in spectroscopy, offering a platform for more efficient and precise analytical techniques.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/716/1*zSPXFiFfpLPpIQ0LsyLU5g.png" /><figcaption>Illustration of how OSLD technology embedded in a smartphone panel may eliminate the need for the screen cutout, or “notch” on the display. Illustration credit: Koala Tech Inc.</figcaption></figure><p>OSLDs are comparable to organic light-emitting diodes (OLEDs), which are ubiquitous in digital displays across various devices. Like OLEDs, organic laser diodes are constructed by sandwiching a thin film of organic material between an anode and a cathode, emitting light when electricity is applied.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tt2yv80rBqlWMi8VKHwLZg.jpeg" /><figcaption>Photo credit: Koala Tech Inc.</figcaption></figure><p>A primary distinction between OSLDs and OLEDs is that OSLDs emit coherent laser light — or “pure light” of a given and tunable wavelength, phase and color — whereas OLEDs emit incoherent light that does not oscillate at the same frequency or phase. This fundamental difference in light characteristics opens the door for OSLDs to have applications that OLEDs cannot cater to. Here are some reasons and applications where OSLDs have an edge:</p><p><strong>Coherence</strong><br> Because OSLDs emit coherent light, they can be used in applications that require precision and specificity, such as in interferometry, holography, and high-resolution microscopy.</p><p><strong>Directionality<br> </strong>Laser light, due to its coherence and typical mode structure, can be highly directional. This makes OSLDs suitable for applications that require directed beams, like optical communication, and low-level laser therapy.</p><p><strong>Data Transmission<br> </strong>Because of their directionality and narrow spectral output, lasers are preferred in communication systems. OSLDs could find a role in optical data transmission, offering potentially high integration density systems.</p><p><strong>Spectral Purity</strong><br> The spectral output of lasers can be much narrower than that of typical incoherent light sources. This makes OSLDs suitable for applications where specific wavelengths or narrow bands of wavelengths are necessary, like in some chemical or biological sensing technologies.</p><p><strong>Biomedical Applications:<br> </strong>Coherent light sources like OSLDs can be used in medical treatments requiring precision, such as targeted photodynamic therapy, laser eye surgeries, or tissue ablation. While OLEDs have their own array of impressive applications, especially in the realm of displays and lighting, their incoherent light output means they are not suitable for the aforementioned applications where the unique properties of lasers (like coherence, directionality, and spectral purity) are indispensable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hzK7qeSSh-exGn206TeLDA.jpeg" /><figcaption>Photo credit: Koala Tech Inc.</figcaption></figure><p><strong>Overcoming Challenges and Pioneering Advancements in OSLD Technology</strong></p><p>However, realizing the full potential of OSLD technology relies on overcoming challenges that include achieving the necessary lasing process under high current conditions, which causes most organic materials to heat up and break down due to electrical resistance when electricity is injected, limiting their lifetime.</p><p>In 2019, Prof. Adachi’s research group demonstrated the viability of organic semiconductor laser diodes, paving the way for novel laser technologies. This breakthrough addressed a fundamental materials science problem, revolving around the identification and optimization of organic compounds capable of producing laser light under electrical excitation.</p><p><strong>Koala Tech: Leading the Way in OSLD Innovation</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/498/1*PMIB3csVjXipx1i_Elu1og.png" /></figure><p>Koala Tech was established to pioneer and commercialize the world’s first organic semiconductor laser diode, making significant R&amp;D advances in materials, device engineering, and photonics to create structures that support the efficient operation of organic materials in a laser diode setup.</p><p>The company has built a strong patent portfolio, proposed new manufacturing and product applications, and forged partnerships with several world-leading, critical industry players on the road to device commercialization. Benefiting significantly from its proximity to the state-of-the-art laboratory and technical infrastructure at the neighboring Kyushu University Ito campus, the international team in Fukuoka continues to collaborate closely with world-leading global research centers for organic optical devices.</p><p>We are excited to back the international Koala team, led by Algerian born Dr. Fatima Bencheikh, as they work to realize the transformative potential of OSLD technologies, set to become a platform for a range of new applications and devices with vast commercial potential.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=62b482c46809" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in Metalmark]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-metalmark-47f2508e074b?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/47f2508e074b</guid>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[boston-startups]]></category>
            <category><![CDATA[biomimicry]]></category>
            <category><![CDATA[deeptech]]></category>
            <category><![CDATA[hello-tomorrow]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Mon, 29 Mar 2021 18:52:37 GMT</pubDate>
            <atom:updated>2021-03-29T18:52:37.915Z</atom:updated>
            <content:encoded><![CDATA[<p>We are pleased to announce our latest investment, <a href="https://www.metalmark.xyz/">Metalmark Innovations</a>, a Boston-based startup that uses biomimicry — i.e., the emulation of nature to address complex human problems — to solve sustainability challenges stemming from sub-micron scale airborne pollutants, toxins, and particles that cause air pollution. The company, which spun out of Harvard University’s <a href="https://aizenberglab.seas.harvard.edu/">Biomineralization and Biomimetics Lab</a>, takes its name from the Metalmark family of butterflies, as the technology is inspired by the organic polymer microstructures in the insect’s wings.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*UTRdkgSsQhtsHw4IXNyqcw.png" /><figcaption>Image credit: Metalmark Innovations</figcaption></figure><p>We see significant potential in Metalmark´s application of materials science and chemistry to imitate nanostructures found in nature, to create novel catalytic materials that can efficiently eliminate airborne pollutants. Elimination of airborne contaminants without the generation of secondary pollutants is a significant benefit. The materials may also be designed to allow for other properties, including anti-viral functions. Beyond air purification Metalmark’s materials have a wide range of potential applications in the form of multifunctional coatings and catalysts, with use cases in the automotive sector, building and construction materials, aviation/aerospace, and more.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*QS0G6DwSPQynVsFG1whELA.png" /><figcaption>Image credit: Metalmark Innovations</figcaption></figure><p>The Metalmark core team is exceptional, consisting of leading Harvard University scientists with backgrounds from materials design, bio-inspired engineering, chemistry and chemical biology, and R&amp;D. The co-founder and CEO started her career in investment banking and tech and has spent the last 20 years promoting tech-enabled sustainability, entrepreneurship and venture investing. The team is advised by Dr. Joanna Aizenberg, the Amy Smith Berylson Professor of Materials Science and Professor of Chemistry &amp; Chemical Biology of Harvard University. Dr. Aizenberg is a world-renowned expert in bio-inspired engineering, with extensive experience from leading major R&amp;D efforts, and startups. Metalmark has to date secured <a href="https://www.metalmark.xyz/post/category/News">multiple grants and awards</a> for their work, including winning the <a href="https://hello-tomorrow.org/2020/12/03/the-14-winners-of-the-global-challenge-2020/">Smart City Track of Hello Tomorrow 2020</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*sPtdj2GfT-M-1rkjfz6Ymw.png" /><figcaption>Image credit: <a href="https://www.urban-x.com/">Urban-X</a></figcaption></figure><p>Various forms of air pollution remain dangerously high in many parts of the world; <a href="https://www.who.int/news/item/02-05-2018-9-out-of-10-people-worldwide-breathe-polluted-air-but-more-countries-are-taking-action">WHO</a> estimates suggest that as many as 9 out of 10 people inhale air containing excessive levels of pollutants, originating from a wide range of sources. By the same estimates ambient (outdoor) and household air pollution cause 7 million deaths annually, from exposure to (ultra)fine particles that penetrate into the respiratory and cardiovascular systems, inducing heart disease and stroke, cancers, respiratory infections, and other chronic illnesses. The current COVID-19 pandemic has drawn further attention to the risk represented by the airborne transmission of viruses. <a href="https://wwwnc.cdc.gov/eid/article/26/9/20-1806_article">Studies indicate</a> retained infectivity and virion integrity of SARS-CoV-2 — the etiologic agent of the coronavirus disease — for up to 16 hours in respirable-sized aerosols in indoor air, emitted through normal human respiration.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*W7tqJOzi4c9oyh_mXVawkA.png" /><figcaption>Image credit: Metalmark Innovations</figcaption></figure><p>We see a bright future for Metalmark ahead, with application areas in airplanes, buildings, the hospitality sector, vehicles, office and residential buildings, and more. The total available market for HVAC and air purification, coatings, air freshening, and catalysts is valued at several hundred billion dollars; <a href="https://www.technavio.com/report/smart-air-purifier-market-industry-analysis">Technavio</a> anticipates that the smart air purifier market size will grow by USD 3.35bn and record a CAGR of 9% during 2020–2024. The Covid-19 crisis has spurred market needs to offer its clients <a href="https://www.nytimes.com/2020/12/03/travel/air-filtration-hotels-cruises.html">the assurance of safe air quality</a>, including in the hotel, cruise, and airline industries. We look forward to supporting the continued upward trajectory of Metalmark Innovations and their cutting-edge inventions to improve air quality, human health, and sustainability.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=47f2508e074b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in Saga Robotics]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-saga-robotics-8bac0e66535c?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/8bac0e66535c</guid>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[robotics-automation]]></category>
            <category><![CDATA[agritech]]></category>
            <category><![CDATA[agriculture-technology]]></category>
            <category><![CDATA[norway]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Tue, 01 Sep 2020 13:52:51 GMT</pubDate>
            <atom:updated>2020-09-01T15:04:03.376Z</atom:updated>
            <content:encoded><![CDATA[<p>Following the recent <a href="https://sagarobotics.com/a/press-kit/press-release/guziapnc/saga-robotics-raises-eur-9-5-million-in-an-effort-to-address-agricultural-labour-shortages-worldwide-with-their-fleet-of-thorvald-robots">press release,</a> we are pleased to announce our investment in <a href="https://sagarobotics.com/">Saga Robotics</a> and their autonomous <a href="https://sagarobotics.com/pages/thorvald-platform">Thorvald</a> agricultural robots. We see great potential in the technology and are joining other investors such as the Norwegian sovereign climate investment company Nysnø, London-based ADM Capital Europe LLP and Rabo Food &amp; Agri Innovation Fund, the investment arm of the Dutch agricultural bank Rabobank.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/830/1*xtC-0Hv7kJzuZ39U9b8gUw.png" /><figcaption>Image credit: Saga Robotics</figcaption></figure><p>If automation of labour-intensive tasks was not obvious before, the Covid-19 crisis made it even more so. Particularly in food production and farming, we are facing shortages of manual labor. If farms were struggling to find enough workers before the crisis, getting them across borders these days is challenging. Another looming and significantly larger risk is the climate crisis: this will challenge the efficiency and sufficient scale-up of food production, but also entail issues such as land degradation and resistance to pesticides.</p><p>Observing these larger trends, we have been tracking robotics and AI-cases in agriculture for a long time. We are therefore very happy to announce our first investment in a Norwegian and UK-based agricultural robotics company, <a href="https://sagarobotics.com/">Saga Robotics</a>. Saga creates autonomous and modular multipurpose agricultural robots that can operate 24/7 and only need an internet connection, access to power, and docking to work. The first commercial products are autonomous robots for UV light treatment mainly of mildew (which also has proven effects against viruses and bacteria in food production), autonomous harvesting, and several other tools under development. UV treatment is an efficient and environmentally friendly alternative to the use of fungicides. The second module currently on offer is a harvester of soft fruits, which operates fully autonomous with 90% accuracy, works continuously at high speed - even at night - and can check qualities while picking.</p><p>The modular components of the Saga robots can efficiently be integrated and easily configured to most farming environments such as open fields, greenhouses, orchards and tunnels. Its modular approach enables low cost even at small-scale production and efficient and cost-effective upscaling of production. In addition to its modular nature, the multipurpose platform distinguishes Saga from its competitors, potentially serving all farmers’ agri-robotics needs down the road. The positive impact of the robots goes even beyond increased efficiency of food production, as the electrically operated robots do no directly emit CO2, do not compress soils and reduce the overall use of pesticides.</p><p>The team at Saga is impressive, with a mix of agricultural robotics expertise, particularly with the CEO as Professor at the Norwegian University of Life Sciences and University of Lincoln (UK), and strong robotics operations and commercial experience. We are pleased to invest in cooperation with a strong and complementary investor group: the Norwegian climate fund Nysnø, the food production giant ADM, the Dutch Rabobank and successful robotics entrepreneur Jakob Hatteland.</p><p>With the first commercial products launched, the initial traction has been strong in the UK and Norway with increasing interest from the US. The farming-as-a-service business model enables early testing and easy scale-up. Saga can thus readily integrate, and with the continuous expansion of its fleet ensure a full transition to autonomous farming. The agriculture market is quickly adjusting to new autonomous solutions, also with established players seeking to acquire new capabilities in autonomous farming (e.g. <a href="https://www.wired.com/story/why-john-deere-just-spent-dollar305-million-on-a-lettuce-farming-robot/">John Deere’s acquisition of Blue River</a>). The Covid-19 crisis could equally leapfrog the market for fully autonomous agri-robotics by positively changing the mentality towards autonomous work and encourage early adoption.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8bac0e66535c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in Universal Quantum]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-universal-quantum-9e1e41d62b34?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/9e1e41d62b34</guid>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Mon, 15 Jun 2020 07:05:44 GMT</pubDate>
            <atom:updated>2020-06-15T08:08:23.858Z</atom:updated>
            <content:encoded><![CDATA[<p>Following the recent <a href="https://www.prnewswire.co.uk/news-releases/universal-quantum-major-investors-back-disruptive-new-player-in-quantum-computing-race-888076874.html">press release</a>, we are delighted to announce our investment in the <a href="https://universalquantum.com/">Universal Quantum</a> seed round, together with Village Global, Hoxton Ventures, Luminous Ventures, 7 Percent, FoundersX Ventures, among others.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/200/1*eIBDiMm35TE4e0eYyjF2Sg.png" /></figure><p>As the company name implies, Universal Quantum’s goal is to build a large scale universal quantum computer, which may represent one of the grandest engineering challenges of our times. If this goal can be achieved however, the benefit to society and potential business opportunities would be enormous.</p><p>After being among the first investors in <a href="https://medium.com/@PropagatorVC/why-we-invested-in-zapata-computing-c8d572d34892">Zapata Computing</a>, a quantum computing software company, we are now doing an investment in a quantum computing hardware company. In recent years, there have been a multitude of proposed hardware architectures and commercial initiatives to scale up such hardware. Arguably, the approach having received the most attention and funding from incumbent players like IBM and Google, is the one based on superconducting qubits, with Google having recently announced their <a href="https://www.nature.com/articles/s41586-019-1666-5">“quantum supremacy” demonstration</a>. This technology can also be considered one of the most mature, as an ecosystem of infrastructure providers and standardisation has already begun to take form. Universal Quantum takes an alternative, but still technologically mature, design approach using ion trapped qubits. Ion trapped qubits offer several advantages over superconducting qubits, one of them being the ability to operate without the need of cooling to milli-Kelvin temperatures using dilution refrigerators. Other approaches, like topological qubits (Microsoft) or semiconductor qubits are currently less mature technologies and arguably out of scope for even the most patient VC-investor. For a recent review see <a href="https://courses.csail.mit.edu/6.857/2019/files/NAE-report-on-quantum-computing.pdf">Quantum Computing: Progress and Prospects (2019)</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/578/1*uUzAqy2FftlY3r-ULxymcg.jpeg" /><figcaption>Image credit: Universal Quantum</figcaption></figure><p><strong>Technology</strong></p><p>Building a truly scalable quantum computer however, poses several additional technological challenges. The team behind Universal Quantum has recently published a blueprint proposing an architecture building on existing proven technological principles and microfabrication techniques to overcome many of these hurdles; see <a href="https://advances.sciencemag.org/content/advances/3/2/e1601540.full.pdf">Blueprint for a microwave trapped ion quantum computer</a>, based on a modular approach making use of ion transport between different modules. While other companies like IonQ make use of laser-driven single- and multi-qubit gates, Universal Quantum’s approach utilizes global long-wavelength radiation and locally applied magnetic fields, a concept first proposed by <a href="https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.87.257904">Mintert and Wunderlich in 2001</a>. This obviates the need for a very large number (on the order of millions) of individual, precisely aligned laser fields which would be required for gate operation for a large scale quantum computer in the laser based approach. Furthermore, using ion transport between modules rather than photonic interconnects allows for a much faster interaction rate between modules. On top of this, a suitable high error-threshold correction code can be applied.</p><p>This is not to suggest that no further technical challenges remains. However, the approach taken by Universal Quantum and their very strong technical and engineering ability makes us believe that this is one of the teams that can get the job done.</p><p><strong>The Team</strong></p><p>At Propagator Ventures we love to back the most promising teams coming out of academia to build commercial success stories. The founders behind Universal Quantum, Winfried Hensinger and Sebastian Weidt, were a perfect match. We especially liked their engineering approach, being willing to depart from the elegant, but less practical road that might offer better acclaim in academic circles. Before investing we met with the team on several occasions discussing the challenges facing a startup coming out of a research group setting, and we were convinced both by the founders tenacity and passion needed to take on such a formidable challenge. Building the company will very much rely on the ability to hire top talent and Universal Quantum has already made 7 new hires of excellent quality.</p><p><strong>Business Plan</strong></p><p>Although still in its infancy, the market and business potential for quantum computing is growing rapidly (see eg. <a href="https://www.bcg.com/publications/2019/quantum-computers-create-value-when.aspx">BCG: “Where Will Quantum Computers Create Value — and When?”</a> or <a href="https://www.mckinsey.com/business-functions/mckinsey-digital/our-insights/a-game-plan-for-quantum-computing">McKinsey: “A game plan for quantum computing”</a>). It is commonly believed that the timeline for commercialization can be divided into the NISQ (Noisy Intermediate-Scale Quantum) era (3–5 years) and the Broad Quantum Advantage to Full Fault Tolerance era (5 years and beyond), although the terms and time estimates vary. Universal Quantum aims for a large scale error corrected universal quantum computer. To achieve that goal certain design choices and compromises will have to made along the way. There is however good reason to believe that several business opportunities will be present also in the near to intermediate term, as witnessed by the increasing willingness of large corporates to engage with quantum computing companies early. Also, Universal Quantum’s modular approach is a natural fit to scaling up. The applicability of NISQ era devices is still to some extent uncertain, but history shows that the most valuable applications might yet not be apparent in the early days of a completely new computing paradigm (<a href="https://www.scientificamerican.com/article/beyond-quantum-supremacy-the-hunt-for-useful-quantum-computers/">Beyond Quantum Supremacy: The Hunt for Useful Quantum Computers</a>).</p><p>We are exited to embark on this journey with Universal Quantum. And we are certain they have one of the best starting points for achieving their goals. Setting up the Universal Quantum Foundation also testifies to their values and the belief that large scale quantum computers will be a technology that should benefit both businesses and society at large.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9e1e41d62b34" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in modl.ai]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-modl-ai-567ff229c815?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/567ff229c815</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[gaming]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Tue, 04 Feb 2020 14:49:51 GMT</pubDate>
            <atom:updated>2020-02-04T14:49:51.805Z</atom:updated>
            <content:encoded><![CDATA[<p>We are thrilled to <a href="https://modl.ai/news/modl.ai-leading-new-wave-in-AI">announce our investment</a> in <a href="http://modl.ai/">modl.ai </a>— a leading scientific and interdisciplinary team that delivers technology accelerating and automating game development and enhancing player engagement. As deep tech investors, gaming was initially not the most obvious area for us to focus on, but we soon came to realize that the field is closely interrelated to AI/ML development in two ways: 1) simulated/virtual environments are already essential to machine learning and we expect a broadening to more diversified AI/ML application areas, 2) AI/ML has great potential to automate the creation of such simulated environments.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*NZlTv7I-6aYOWmj_Wn-p3Q.png" /></figure><p><strong>Gaming for AI-training</strong></p><p>As there are practical limitations to how much real-world data can be collected for AI-training purposes, an increasing number of AI models rely on simulated data environments. Simulated data can increase both quantity and quality of training sessions to better resemble the complexity of the real-world. Especially when it comes to understanding and modeling of human behavior, the need for diverse scenarios and oftentimes unpredictable behavior significantly increases complexity. (Synthetic data creation as in replicating and adjusting existing datasets does not necessarily allow for the same complexity and diversity as a fully simulated environment.)</p><p>Games have become an especially potent training environment as it increasingly embeds multiple and interacting players and highly complex contexts and environments as well as edge cases. Alphabet’s Deepmind has recently made <a href="https://deepmind.com/blog/article/AlphaStar-Grandmaster-level-in-StarCraft-II-using-multi-agent-reinforcement-learning">strong advances in adapting AI agents to complex games such as StarCraft</a>; other researchers <a href="https://science.sciencemag.org/content/365/6456/885">showed lately that AI agents can navigate multiplayer games such as six player poker</a>. Particularly in areas such as self-driving cars (UAVs) have we seen the application of <a href="https://www.nytimes.com/reuters/2020/01/06/business/06reuters-tech-ces-amazon-com.html">test environments that can model all kinds of unpredictable human behavior before implemented in real life</a>.</p><p>Also in the area of simulating data for VR/AR environments, we see interesting developments like FBs recently launched <a href="https://ai.facebook.com/blog/open-sourcing-ai-habitat-an-simulation-platform-for-embodied-ai-research/">AI Habitat</a>, which offers an advanced and open source simulation platform for embodiment of agents in 3D life-like environments.</p><p>As these initial technological developments indicate, we expect virtual environments created by gaming (or VR) developers to become an increasingly important source for future AI training.</p><p><strong>Automating gaming development</strong></p><p>Now, complex gaming environments are themselves laborious to create. Especially demanding tasks are manual game testing, continuous game updating of content and code, balancing of games, anomaly detection and toxic player/bot detection. Beyond high time consumption, manual testing and one-off game development may lead to shallow user models and lacking personalization which again result in low retention, missing revenue potential and poor player experiences. Freeing up game developers’ time normally spent for such tasks, may further create room for more productive creative design.</p><p>modl.ai creates AI personas based on models of actions, motivations and experiences. By combining psychometric insights and AI methods they can predict what users will do, want and experience more accurately and faster than current models. Thus, based on player profiling they can evaluate content based on personas, create game-playing agents, detect anomalies and personalize content, all in an autonomous fashion.</p><p><strong>Games as a service &amp; beyond</strong></p><p>The gaming market in itself is a huge and lucrative market (TAM 150bn in 2019; expected to grow to 300bn by 2025). Games as a service, or GaaS, is a massive opportunity and a trend expecting to substantially overtake more of the gaming market in the near future. As games as a service require continuous adaptation and new iterations every couple of weeks, the trend is making the gaming industry go from one-off products to creating a long-term relationship with their players.</p><p>Short update cycles may strain production resources and fast updates risk breaking existing content. Thus, continuous evaluation of performance calls for scaling across content and increased focus on life-time of players. Game performance is more likely to have periodic peaks in performance and traction instead of a one-time high with a flattening-out over time.</p><p>The GaaS trend enforces the need for automating game development and strengthens the business case of modl.ai. Better prediction of user behavior and personalized content will increase life-time performance and help build the long-term relationship with players. Insights on player profiles and motivation will also offer valuable information to game developers before, during and after the game development process. More accurate profiles will increase player engagement and allow for content adaptation to personas. It will also enable detection of toxic or bot players in games, particularly relevant for monetized incentivisation.</p><p>This automated and procedural content generation, defined as “the algorithmic creation of game content with limited or indirect user input” (<a href="https://link.springer.com/chapter/10.1007/978-3-319-42716-4_1">Togelius et al., 2016</a>), further goes hand in hand with a recurring revenue model, having the potential to generate substantial and more stable revenue for game developers and the tools they use (such as modl.ai).</p><p><strong>Team:</strong></p><p>We believe modl.ai with the team’s strong academic record and multidisciplinary approach will change how gaming simulations are automated in a human-like fashion, something the first commercial traction and great user feedback testify to. They are an academically accomplished team combining expertise from areas of AI/ML for gaming, gaming industry, psychology and UX. With 10+ PhDs on the team and some of the most highly published researchers in AI gaming (e.g. from New York University, University of Malta, IT University of Copenhagen), the team’s diverse capabilities make them stand out. Next to various highly cited articles on procedural content generation in gaming and player modeling, the co-founders Julian Togelius and Georgios N. Yannakakis have for example recently published a <a href="https://www.springer.com/gp/book/9783319635187">textbook on AI in gaming (2018)</a>. On top of this, the team has a long history of professional game development, having launched multi-platform and critically acclaimed game titles, giving them first-hand insight into the needs of game developers.</p><p><strong>Modeling human behavior:</strong></p><p>The market for automating game development is interesting in itself, but also for gaming as simulation environments as it expands to more human-realistic simulations beyond the gaming industry. As AI training data demand expands beyond UAVs to include more complex human behavior, we expect the simulations market to grow in parallel direction.</p><p>Being able to model complex 3D games and adjust simulations in real-time will give the autonomous approach of modl.ai tremendous potential in various applications for simulations. To realistically model human behavior and offer personalized simulation environments will demand an interdisciplinary approach, and we believe this team is the right fit for the job.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=567ff229c815" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in Kebotix]]></title>
            <link>https://propagatorvc.medium.com/why-we-invested-in-kebotix-5568f2516f1a?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/5568f2516f1a</guid>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Thu, 08 Nov 2018 10:31:40 GMT</pubDate>
            <atom:updated>2020-04-05T13:22:49.015Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/512/1*defBatasUPDixQWppkBYKw.png" /></figure><p>Following the official <a href="https://www.kebotix.com/press/2018/11/7/harvard-scientists-launch-breakthrough-ai-and-robotics-tech-company-for-rapid-innovation-of-materials">press release</a> of the seed-round and the great coverage in <a href="https://www.technologyreview.com/s/612388/a-robot-scientist-will-dream-up-new-materials-to-advance-computing-and-fight-pollution/">MIT Tech Review</a>, we are proudly announcing our second investment and highly promising addition to our portfolio: Kebotix.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZTAU5nM7KviV6_2n5j3V3A.png" /></figure><p><strong>What they do</strong></p><p>Kebotix has created a fully autonomous material science discovery lab. It combines machine learning algorithms to model molecular structures with an autonomous robotics lab that synthesizes, tests and feeds back the results to the algorithms. By applying deep learning, the algorithms adapt to the results from the lab, creating a ‘closed loop’ for fast learning and simulation. In fact, an early demonstration of the tech platform, which was sold before Kebotix was incubated, was able to reduce the development time of OLED materials from previously 7 to 1.5 years<strong>. The platform thus shows immense potential for developing new green materials which our world desperately needs fast</strong> (e.g. <a href="https://www.nature.com/news/can-artificial-intelligence-create-the-next-wonder-material-1.19850">Nature article</a>)!</p><p><strong>First step: Molecular Inverse-Design</strong></p><p>The discovery process of materials is mostly still following a heuristic, step-by-step procedure: The researcher has some pre-defined hypotheses or intuition about certain combination on a molecular level. What follows is limited testing of those specific molecular combinations with a discovery cycle normally stretching over several years. Considering that the size of chemical space — all possible chemically active molecular combinations — is in the order of 10⁶⁰ molecules, the task of mapping all such combinations seems too exhaustive to solve with the current (human) approach.</p><p>By using a technology named inverse design, Kebotix sets up their algorithms to look for optimal molecular combinations based on desired properties (e.g. toxicity, solubility, activity). The ‘inverse’ connotation denotes the difference from the traditional heuristic approach, which would first test these properties after the process of conceptualizing and synthesizing such combinations. With the inverse process, the machine learning approach performs much more systematic searches across the combinatorial space and with optimizing strategies are able to simulate likely molecular properties before physical testing (see figure below). For the more techy ones, Kebotix uses several leading machine learning techniques: Variational Auto Encoders (VAEs), Generative Adversarial Networks (GANs), and reinforcement learning. For more details, this recent <a href="http://science.sciencemag.org/content/361/6400/360">Science article</a> explains the methods and process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*iMfstf4_hRhq2rZlCY_WGA.png" /><figcaption>Source: <a href="http://science.sciencemag.org/content/361/6400/360">Science 2018</a></figcaption></figure><p><strong>Second step: Robotics</strong></p><p>While one has seen several inverse-design propositions within drug discovery (notable startups are BenevolentBio, Zymergen, Atomwise, a.o.), there are only very few trying to do the same for materials science. There has been some advances on <em>automating</em> the vet lab testing. <strong>But none have yet to develop an <em>autonomous </em>or ‘self-driving’-lab in materials science</strong> that exploits the closed learning loop described above. Several of the co-founders of Kebotix have developed an open source architecture (‘ChemOS’) to run and synchronize such autonomous labs with machine learning algorithms. Thus<strong>, the field is moving away from automated to autonomous discovery, which is facilitating novel discoveries at an increasing pace</strong>. For a description of the ChemOS software for autonomous labs, see this recent <a href="https://chemrxiv.org/articles/ChemOS_An_Orchestration_Software_to_Democratize_Autonomous_Discovery/5953606">research article</a>.</p><p><strong>Business-model</strong></p><p>The closed loop model is unique within materials science prediction and thus has a clear commercialization potential: Chemical, pharmaceutical and other materials producers on molecular level are all interested in reducing time to market in R&amp;D. R&amp;D spending of chemical materials is around $50bn yearly — while the software spending remains a very low share (&lt;1%), with innovation in the discovery process towards algorithms and machine learning as seen in drug discovery, the market is likely to massively increase in the next years.</p><p>Kebotix offers external companies access to its ‘ChemOS’ as a service, while ensuring anonymity of data. One could say, that <strong>while Labster is making virtual learning in labs more easily accessible, Kebotix’ remotely accessible lab is making actual experiments in a world-leading lab environment accessible</strong>. The value add by more efficient and faster time to discovery is obvious to both commercial clients and society (we need to speed up development of green materials, better batteries, degradables, etc.). Further, the revenue model of recurrently renting out access to a platform with relatively minor scalability issues (only the remote robotics labs will need to increase according to demand) is highly attractive.</p><p>Considering the longer term, however, it could also make sense for Kebotix to directly take advantage of the possible massive IP generation of such material discovery potential in its lab — they could by themselves develop novel materials inhouse to license and sell directly.</p><p>Further, <strong>the materials science discovery market is up for grabs</strong>. While we’ve seen highly innovative models in drug discovery (BenevolentBio, Zymergen, a.o.), materials science have largely remained overlooked. Kebotix, with its strong team, is more than capable to take a leading position here.</p><blockquote><strong>We are about to surf a new wave of innovation in material science. Don’t watch from the beach, grab a board and paddle out… (Kebotix)</strong></blockquote><p><strong>The Team</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4TSPWR1sLm_WJ9o5qiRY1w.jpeg" /><figcaption>Team of Kebotix (left to right): Prof. Alán Aspuru-Guzik, Dr. Dennis Sheberla, Dr. Jill Becker, Dr. Semion Saikin, Dr. Christoph Kreisbeck.</figcaption></figure><p>The team is comprised of top-notch researchers from Harvard within chemistry, physics, machine learning and CS. (We are very fond of such interdisciplinary teams tackling a problem together from diverse perspectives!) The research group from which the team spun-off has long been <a href="https://www.technologyreview.com/s/602756/software-dreams-up-new-molecules-in-quest-for-wonder-drugs/">considered to be leading</a> within molecular chemistry. Jill, CEO with a PhD in chemistry and a previous entrepreneur, has a <strong>rare combination of technical wit and commercial intuition</strong>, which immediately convinced us the first time we encountered her.</p><p>We were the first investor in the company and are thrilled to continue observing the great work of the team. The latest proof being the closing of an oversubscribed round in July with investors such as Baidu Ventures, One Way Ventures &amp; Flybridge Capital. Considering the strong impact potential of the platform with regards to new materials, this one should be on everyone’s radar!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5568f2516f1a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Investment Hypotheses]]></title>
            <link>https://propagatorvc.medium.com/investment-hypotheses-c7cb369b821?source=rss-ab3eeb5caf3d------2</link>
            <guid isPermaLink="false">https://medium.com/p/c7cb369b821</guid>
            <category><![CDATA[robotics]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[quantum-computing]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Propagator Ventures]]></dc:creator>
            <pubDate>Sun, 09 Sep 2018 14:48:53 GMT</pubDate>
            <atom:updated>2018-09-11T14:17:59.299Z</atom:updated>
            <content:encoded><![CDATA[<p>Propagator Ventures positions itself early in the technology trajectory. We keep in close contact with our research and scientific network and seek to identify emerging technology trends on the brink of commercialization and venture capital attention.</p><p>Our hypotheses are based on the convergence of several general trends: machine &amp; deep learning advances, engineering and computing at the nanometer level, and the commoditization of hardware and sensors. These trends often overlap and are observed in combination (e.g. Kebotix, combining new machine learning with robotics for materials discovery).</p><p>While our hypotheses remain under constant development (just like in science), we have listed several hypotheses below that describe which investment opportunities in emerging tech that we currently see:</p><h4>1. Machine learning/AI</h4><p>While large tech players dominate algorithm development and machine learning models, we see opportunities within more specialized domains requiring contextualized knowledge and access to data:</p><ul><li><strong>Health</strong> data is often proprietary, private and not centralized in hospitals &amp; clinics. Further, advances in genomics, proteomics and biomics as well as new imaging techniques &amp; sensor data offer rich datasets. Expected opportunities within decentralized data approaches, encryption and deep learning.</li><li><strong>Heavy-asset industries </strong>are faced with continuous monitoring and complex collection of fragmented and decentralized data. New algorithm development has shown potential in handling such complex systems in a smarter way and to better predict critical system failure.</li></ul><h4>2. Quantum computing</h4><p>Breakthroughs in quantum computing technology are rapidly progressing together with first-movers and expert teams. There is a window of opportunity for investing in such founders today before larger players or consolidation emerge.</p><h4>3. Robotics</h4><p>The rapid development in robotics is driven by commoditization of hardware and sensors and innovations in control algorithms, computer vision and deep learning. While AGI is still a distant goal, new highly skilled robotics systems has the potential to transform several industries in the near to intermediate term.</p><h4>4. Encryption</h4><p>Quantum computing may render current encryption standards obsolete within 10–15 years necessitating the need for wide deployment of quantum safe encryption algorithms. Increasing focus on privacy and ownership of data from governments and individuals will produce significant opportunities for new encryption technology.</p><h4>5. New materials</h4><p>The space of chemically active molecular combinations is huge (10⁶⁰) and as traditional materials discovery is guided by heuristics and expert knowledge, the process is slow. New machine learning techniques together with the ability to engineer at the nanometer scale will create opportunities for accelerated discovery of new materials through scalable business models.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c7cb369b821" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>