<?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[TPY Capital - Medium]]></title>
        <description><![CDATA[We invest in companies that leverage the exponentially growing power of data and machine intelligence across multiple verticals as well as those who make enterprise organizations more efficient and effective through productivity tools. - Medium]]></description>
        <link>https://medium.com/tpy-capital?source=rss----9eff3e2c8cf2---4</link>
        <image>
            <url>https://cdn-images-1.medium.com/proxy/1*TGH72Nnw24QL3iV9IOm4VA.png</url>
            <title>TPY Capital - Medium</title>
            <link>https://medium.com/tpy-capital?source=rss----9eff3e2c8cf2---4</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 20:25:34 GMT</lastBuildDate>
        <atom:link href="https://medium.com/feed/tpy-capital" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[9 predictions for GenAI in 2026, and how our 2025 calls actually played out]]></title>
            <link>https://medium.com/tpy-capital/9-predictions-for-genai-in-2026-and-how-our-2025-calls-actually-played-out-a2caa442fbba?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/a2caa442fbba</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[quantum-computing]]></category>
            <category><![CDATA[predictions]]></category>
            <category><![CDATA[genai]]></category>
            <dc:creator><![CDATA[Amit Sabag]]></dc:creator>
            <pubDate>Tue, 20 Jan 2026 11:59:35 GMT</pubDate>
            <atom:updated>2026-01-20T11:59:35.058Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*HVGRElo4nORdnwvBH94YhQ.png" /></figure><p>Making predictions is uncomfortable. If they all come true, you probably stayed too close to consensus. If nothing comes true, you were just guessing (or maybe downright stupid?!). The only honest place to operate is the frontier, where the signals are noisy, the outcomes are uncertain, and being wrong is part of the process.</p><p>That is why we publish predictions publicly. They force us to sharpen thousands of signals into a coherent worldview. They also make us accountable. This post does both. First, nine concrete predictions for 2026. Then, a quick look back at our 2025 predictions, where we were directionally right, where we missed, and what we learned.</p><h3>Part I: 9 predictions for 2026</h3><h4>1. Fortune 500 companies have reached peak employment; and from now on it is decline</h4><p>2026 marks the first sustained decline in headcount among the largest enterprises. This is not global unemployment collapse, but a clear signal that productivity gains from AI are now large enough to change organizational math.</p><p>AI no longer augments at the margins. It moves the productivity curve. The break between job openings and equity performance is the tell. <strong>White-collar layoffs in 2025 were not a cycle. They were a preview.</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*79MYvvUliFGl6yjS.png" /></figure><h4>2. Quantum, AI, and their convergence will be a major theme in 2026</h4><p>2026 is when quantum computing stops being discussed in isolation. The combination of quantum and AI enters the core narrative, driven by real progress in hybrid architectures.</p><p>The opportunity spans cryptography, materials, drug discovery, and optimization. Infrastructure players are already preparing for this convergence (e.g., <a href="https://www.nvidia.com/en-eu/solutions/quantum-computing/nvqlink/">NVQ link</a> by Nvidia).</p><h4>3. The next trillion-dollar company will emerge from the AI memory industry</h4><p>Compute gets the headlines. Memory is the bottleneck. Training and inference are increasingly constrained by memory bandwidth and availability. The dominant memory suppliers already face demand they cannot satisfy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*eRQlBO3_bSp17Ao-.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*2Ek_jWPox1KLYknb.png" /></figure><h4>4. Regulators crack down on circular revenue</h4><p>Accounting scrutiny increases around partnerships, credits, investments, and commercial deals where money circulates among the same players. Revenue quality will matter more than revenue growth.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*-oOkAHvhvtmFS-PG.png" /></figure><p>At the same time, depreciation on AI infrastructure will be reassessed, revealing higher real costs than many models assume. Valuations will feel the pressure.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*GfOGxO5D5g4tCvdB.png" /></figure><h4>5. Anthropic will begin developing its own hardware to gain a competitive advantage</h4><p>2025 was the year when everyone dances with everyone.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*t4WU-UgX5M1Hpgxw.png" /></figure><p>Large AI labs cannot remain purely software companies since competitive advantage increasingly lives lower in the stack. Anthropic will start developing its own hardware components, following the path already taken by other hyperscalers. This expands its dependency graph across semiconductor suppliers and deepens the hardware race.</p><h4>6. 1% of all global eCommerce transactions will be agent driven</h4><p>We predict that by the end of 2026, one percent of global eCommerce transactions will be executed by AI agents. Today, the number is closer to 0.001%.</p><p>This triggers a cascade:</p><ul><li>Commerce designed for agents, not humans (e.g., <a href="https://newsroom.paypal-corp.com/2025-10-28-OpenAI-and-PayPal-Team-Up-to-Power-Instant-Checkout-and-Agentic-Commerce-in-ChatGPT">OpenAI and PayPal partnership</a>)</li><li>AEO replacing SEO</li><li>Agent-to-agent protocols for negotiation, verification, and settlement (e.g., <a href="https://stripe.com/blog/agentic-commerce-suite">Stripe’s ACP protocol</a>)</li></ul><p><strong>The deeper shift is from navigation to delegation. The browser stops being a place you work in. It becomes a system that works for you.</strong></p><h4>7. Entertainment becomes real-time and personal</h4><p>You no longer follow Season 4 Episode 7. You follow your canonical version of the series, which updates itself as if the writers’ room sits inside your profile. Stories adapt in real time to mood, preference, and context. We began to see, in 2025, <a href="https://z.ai/blog/realvideo">the emergence of technologies </a>capable of enabling real-time video creation. This will reshape studios, creators, and IP economics.</p><h4>8. Global breach exposes millions of personal embeddings</h4><p>One global breach exposes tens of millions of records and makes AI security a board-level and national concern. The response mirrors past security shifts: more segmentation, more local execution, stronger embedded models on devices. We already saw an early preview of this in the <a href="https://openai.com/index/mixpanel-incident/">OpenAI Mixpanel incident</a>.</p><h4>9. Google is the new leader — paradoxically, this could be the worst relative position Google will ever be in</h4><p>We said it last year as well, it was far too early to dismiss Google. Now is the moment to expect an even bigger future, one that will become clear in 2026. What Google has accomplished over the past six months is genuinely extraordinary. Near-total dominance in video (<a href="https://aistudio.google.com/models/veo-3">Veo 3</a>) and image generation (<a href="https://gemini.google/overview/image-generation/">Nano Banana</a>), with products that lead the market. Strength in text as well. <a href="https://www.theverge.com/report/827555/google-gemini-3-is-winning-the-ai-race-for-now">Gemini 3 currently sits at the top of the benchmarks</a>.</p><p>Beyond benchmarks, user engagement metrics also place Google in the lead, and the same holds true for download volumes, where Google has effectively caught up with OpenAI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*iJ8RSYvoM8D2yyrR.png" /></figure><p>All of this reflects the massive assets Google brings to the table, distribution, deep talent pools, and unparalleled data, a combination that is extremely difficult to compete with.</p><h3>So to summarize..</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*s6_iiRnb4jH_NHw5.png" /></figure><h3>Part II: A quick scorecard for 2025</h3><h4>1. Fortune 500 companies will have an average of three AI agents in production</h4><p>We predicted that Fortune 500 companies would run multiple AI agents in production. Reality landed slightly below that bar, but the inflection clearly happened. Despite growing skepticism around ROI, data tells a different story. By late 2025, a significant portion of US enterprises were already paying for AI licenses. Production friction is real, but adoption momentum is stronger.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*-pEtBY-6gAVzv4CG.png" /></figure><p><strong>Verdict:</strong> Right direction, optimistic timing.</p><h4>2. The first one-person company valued at $1b</h4><p>We predicted a one-person unicorn. That exact milestone has not materialized, but the underlying dynamic became undeniable. Native AI companies are generating unprecedented revenue per employee. Small teams reached massive valuations at speeds that would have been unthinkable two years ago.</p><p><strong>Verdict:</strong> Missed the headline, nailed the structural shift.</p><h4>3. Companies with proprietary datasets will generate +30% of revenue from data deals</h4><p>We expected data owners to begin extracting meaningful value through licensing. While the 30% threshold turned out to be aggressive, 2025 marked the normalization of large-scale data deals, including <a href="https://searchengineland.com/openai-may-pay-reddit-70m-for-licensing-deal-451882">OpenAI with Reddit</a>, <a href="https://openai.com/index/openai-and-guardian-media-group-launch-content-partnership/">OpenAI with Guardian Media</a>, and <a href="https://openai.com/index/disney-sora-agreement/">OpenAI with Disney</a>. At the same time, hyperscalers started actively exploring solutions for the long tail of content owners, for example <a href="https://techcrunch.com/2025/03/21/microsoft-is-exploring-a-way-to-credit-contributors-to-ai-training-data/">Microsoft examining mechanisms to credit contributors whose data is used in AI training</a>.</p><p><strong>Verdict:</strong> Directionally correct.</p><h4>4. Developing countries will experience the fastest GenAI adoption rates</h4><p>Adoption remained led by developed economies.</p><p><strong>Verdict:</strong> Wrong.</p><h4>5. OpenAI’s market share in the LLM space drops by 50%</h4><p>We predicted a sharp drop in OpenAI’s market share. The decline happened, though slower than anticipated. Competitive pressure intensified, especially from Google and new model entrants. The direction is clear, even if the endpoint takes longer.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*5CIW19p8gS8HgL7W.png" /></figure><p><strong>Verdict:</strong> Early, but on track.</p><h4>6. The beginning of the end…</h4><p>In 2025, we predicted that we would witness the end of two eras.</p><h4>6A. The beginning of the end for mobile devices</h4><p>On one hand, 2025 exposed how hard it is to replace the smartphone. Early AI-first devices struggled to find product market fit. <a href="https://www.theverge.com/news/614883/humane-ai-hp-acquisition-pin-shutdown">Humane shut down</a> the AI Pin and sold its remaining assets, <a href="https://techcrunch.com/2025/07/22/amazon-acquires-bee-the-ai-wearable-that-records-everything-you-say/">Amazon quietly acquired Bee</a>, another wearable AI device company, for an undisclosed sum.</p><p>On the other hand, the long-term direction became impossible to ignore. <a href="https://www.cnbctv18.com/technology/samsung-smart-glasses-haean-project-moohan-xr-meta-ray-ban-apple-vision-pro-19578420.htm">Samsung unveiled new smart glasses</a>. <a href="https://www.meta.com/blog/project-aria-gen-2-next-generation-egocentric-research-glasses-reality-labs-ai-robotics/?srsltid=AfmBOoqL1GxRD3hSOrhy_eMJ9AErG4KYi052bjzl5QD4v-52zH_AtNRZ">Meta announced its Aria Gen 2 research glasses</a>. <a href="https://techcrunch.com/2025/05/21/jony-ive-to-lead-openais-design-work-following-6-5b-acquisition-of-his-company/">OpenAI acquired Jony Ive’s company for $6.5b</a>. This was not an acquihire or a feature play. It was a strategic bet on creating an entirely new computing platform.</p><p>First-generation attempts are failing loudly, but platform builders are doubling down. The smartphone is not disappearing tomorrow, but 2025 marked the moment the industry collectively accepted that its dominance is no longer guaranteed.</p><p><strong>Verdict: </strong>Directionally correct.</p><h4>6B. The beginning of the end of search engines as we know them</h4><p>In 2025, the world of search was fundamentally reshaped. The shift began with the rapid emergence of deep research capabilities (<a href="https://openai.com/index/introducing-deep-research/">OpenAI</a>, <a href="https://www.perplexity.ai/hub/blog/introducing-perplexity-deep-research">Perplexity</a>, <a href="https://gemini.google/overview/deep-research/">Google</a>, and many more…). Search stopped being about retrieving links and started becoming about synthesizing knowledge, reasoning across sources, and generating structured insights in real time.</p><p>Next came AI-native browsers (<a href="https://openai.com/index/introducing-chatgpt-atlas/">Atlas </a>and <a href="https://www.perplexity.ai/comet">Comet</a>). Rather than acting as passive windows into the web, they orchestrated searches across multiple modalities and embedded results directly into user workflows, even if the experience still feels clunky and early.</p><p>Finally, Google’s launch of <a href="https://search.google/ways-to-search/ai-mode/">AI Mode</a> marked a clear inflection point. By integrating generative and reasoning capabilities directly into the core search experience, Google effectively signaled that traditional search paradigms were no longer sufficient.</p><p><strong>Verdict: </strong>Correct.</p><h4>7. The launch of the first household robot</h4><p><a href="https://fortune.com/2025/10/30/1x-neo-household-robot-chores-20k/">1X unveiled NEO with a $20K price point</a>. While the announcement sparked criticism, with claims that the robot was partially operated remotely, the category still made a meaningful leap forward. Consumer robotics crossed an important threshold in visibility, ambition, and perceived feasibility.</p><p><strong>Verdict:</strong> Correct.</p><h4><strong>8. Nvidia’s share performance will outperform all other hyperscalers</strong></h4><p>We were right until November, and then Google started to lead (a spoiler for one of our 2026 predictions).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*WqYjNRmyiLGjTwsl.png" /></figure><p><strong>Verdict:</strong> Almost.</p><h4><strong>9. Data centers fail to meet energy demands</strong></h4><p>Data centers sat idle because electricity could not reach them. This went from fringe concern to core bottleneck.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*rZGiVVJ10rQ1gwOg.png" /></figure><p><strong>Verdict:</strong> Correct</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/875/0*l_yTdLzws_wtPASl.png" /></figure><h3>Keep us honest</h3><p>Predictions are only useful if they are revisited. This is our stake in the ground for 2026. We will revisit this. See you in 2027😊</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a2caa442fbba" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/9-predictions-for-genai-in-2026-and-how-our-2025-calls-actually-played-out-a2caa442fbba">9 predictions for GenAI in 2026, and how our 2025 calls actually played out</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[9 bold predictions for GenAI in 2025]]></title>
            <link>https://medium.com/tpy-capital/9-bold-predictions-for-genai-in-2025-ad656214b4ab?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/ad656214b4ab</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[predictions]]></category>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[genai]]></category>
            <category><![CDATA[venture-capital]]></category>
            <dc:creator><![CDATA[Amit Sabag]]></dc:creator>
            <pubDate>Wed, 22 Jan 2025 14:10:10 GMT</pubDate>
            <atom:updated>2025-01-22T14:10:10.252Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*RaxAv64EKtJVOkB2" /></figure><p>As 2024 draws to a close, it’s clear that GenAI is one of the most transformative revolutions we’d witnessed. In this special episode of the <a href="https://www.youtube.com/@Aideation-TPY/videos"><em>Aideation</em> podcast</a>, hosted by <a href="https://www.tpycapital.com/">TPY Capital</a>, we gave some pretty bold predictions for where GenAI is headed in 2025. Here’s a concise breakdown of the trends and shifts we believe will define this year:</p><h3><strong>1. By the end of 2025, Fortune500 companies will have an average of three AI agents in production</strong></h3><p>These agents — autonomous systems capable of executing complex, end-to-end processes — will move beyond the experimental phase (POC) into scalable production.</p><p><strong>Why It Matters:</strong></p><ul><li>Today, most use cases are stuck in the POC stage. Scaling AI agents represents a major engineering challenge akin to the “productization” of GenAI.</li><li>Key adoption areas will include code generation, marketing, customer support, and sales — domains where technology is becoming mature, adoption risks are relatively low, and the financial gains are high.</li><li>Adopting enterprises will require robust infrastructure for agent authentication, integration, and runtime governance.</li></ul><p><strong>The Big Question:</strong> Who will control these agents? Internal teams, LLM providers, vertical-specific startups. The battle for dominance is just beginning.</p><h3><strong>2. 2025 could see the rise of the first one-person company valued at $1 billion</strong></h3><p>GenAI’s democratization of product development makes this provocative prediction more plausible than ever.</p><p><strong>What’s Driving It:</strong></p><ul><li>AI agents and automation tools drastically lower the barriers to building and scaling companies.</li><li>NVIDIA’s CEO hinted of a future where companies grow revenue exponentially without proportional increases in headcount.</li></ul><p><strong>Implications for VCs:</strong> This shift challenges traditional SaaS and startup funding models. As solopreneurs quickly identify market arbitrage opportunities and achieve profitability, the role of venture capital will require its own adaptations.</p><h3><strong>3. Companies that own proprietary datasets (text, images, video, or code) will generate at least 30% of their revenue from data licensing deals</strong></h3><p><strong>Key Trends:</strong></p><ul><li>Data scarcity is becoming a bottleneck for AI production. Licensing agreements, such as Reddit’s $60M deal with OpenAI, highlight this emerging revenue stream.</li><li>The next wave will involve multi-modal data (e.g., media companies like Netflix and Disney licensing video datasets).</li><li>Regulation around content ownership and user data rights will shape this trend.</li></ul><h3><strong>4. Developing countries will experience the fastest GenAI adoption rates in 2025</strong>, driven by:</h3><ul><li>Low-cost AI models tailored for specific tasks (SLMs).</li><li>Edge AI enabling on-device processing without pre-training.</li><li>Infrastructure expansion, such as new data centers in Africa and Asia.</li></ul><p><strong>Historical Context:</strong> Similar to the mobile and fintech revolutions, these regions could leapfrog traditional platforms established in more developed locales, adopting cutting-edge technology at scale.</p><h3><strong>5. 2025 will mark the beginning of the end for mobile devices and search engines as we know them</strong></h3><p><strong>Mobile:</strong></p><ul><li>Emerging technologies like wearable AI (e.g., Meta’s Ray-Ban glasses) could challenge the smartphone’s dominance.</li><li>Voice-to-voice AI models are on track to pass the Turing test, enabling seamless, conversational interactions.</li></ul><p><strong>Search:</strong></p><ul><li>Conversational interfaces are rapidly replacing traditional search engines. By 2025, over 30% of search interactions will be chat-based.</li><li>Google’s homepage is expected to undergo its first major redesign in over a decade to reflect this paradigm shift.</li></ul><h3><strong>6. 2025 will see the launch of the first household robot designed for tasks like cleaning, organizing, and grocery management</strong></h3><p><strong>Why Now?</strong></p><ul><li>Advances in robotics and language models are converging to enable this long-awaited consumer breakthrough.</li><li>Major players like Tesla, Boston Dynamics, and Figure AI are leading the charge.</li><li>This could be the most significant consumer revolution since the mobile era of 2008.</li></ul><h3><strong>7. OpenAI’s dominance in the LLM space will erode significantly, with its market share dropping by 50%</strong></h3><p><strong>Why?</strong></p><ul><li>Open-source models offer more control and lower costs, appealing to enterprises.</li><li>Competitors like Google (leveraging unique data assets) and Anthropic are closing the performance gap.</li><li>OpenAI’s reliance on Microsoft adds strategic vulnerability, as Microsoft diversifies its AI bets.</li></ul><h3><strong>8. In 2025 Nvidia share performance will surpass that of any other hyperscalers</strong></h3><p>NVIDIA had been the undisputed winner of this revolution and had already generated jaw dropping returns, but in 2025 its share performance will surpass that of any other hyperscalers, including like AWS, Google, Microsoft and Meta.</p><p><strong>Why?</strong></p><ul><li>Reasoning models’ GPU requirements will prove to be a boon for Nvidia.</li><li>More tokens run, with a cost per token which is higher, will feed into Nvidia’s performance.</li></ul><h3><strong>9. For the first time, data centers may fail to meet energy demands due to GenAI’s exponential compute requirements</strong></h3><p><strong>Supporting Stats:</strong></p><ul><li>A single ChatGPT query consumes 10x more energy than a Google search.</li><li>Global data center electricity usage is expected to double by 2030, posing serious sustainability challenges.</li><li>Potential solutions include nuclear energy and strategic investments in renewable sources.</li></ul><p><strong>Looking Ahead</strong></p><p>These predictions are a drop in the bucket of trends GenAI will go through in 2025. Whether it’s autonomous agents reshaping workflows, the democratization of entrepreneurship, or the reinvention of consumer devices, one thing is certain: the future of AI is arriving faster than anyone could have imagined…not even twelve months ago!</p><p>Stay tuned as we revisit these forecasts in a year’s time to see how close we came to predicting the unpredictable.</p><p><em>Thank you for joining us on this journey! Don’t forget to follow the </em><a href="https://www.youtube.com/@Aideation-TPY/videos"><em>Aideation podcast</em></a><em> for more insights and conversations at the forefront of the GenAI revolution.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ad656214b4ab" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/9-bold-predictions-for-genai-in-2025-ad656214b4ab">9 bold predictions for GenAI in 2025</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why We Invested in Datomize]]></title>
            <link>https://medium.com/tpy-capital/why-we-invested-in-datomize-d8ceb20e1c4?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/d8ceb20e1c4</guid>
            <category><![CDATA[fintech]]></category>
            <category><![CDATA[synthetic-data]]></category>
            <category><![CDATA[gdpr-compliance]]></category>
            <category><![CDATA[data-collection]]></category>
            <category><![CDATA[privacy]]></category>
            <dc:creator><![CDATA[Dekel Persi]]></dc:creator>
            <pubDate>Wed, 24 Feb 2021 16:40:07 GMT</pubDate>
            <atom:updated>2021-02-24T16:40:07.833Z</atom:updated>
            <content:encoded><![CDATA[<p><a href="http://www.datomize.com/">Because privacy matters</a></p><p><strong>For several years, TPY has had a strong conviction about the future of enterprise data. We believe that the volume and complexity of data will continue to increase exponentially. That this data can and should be used to expand business activity. And that currently, doing so could pose privacy and compliance challenges.</strong></p><p>As a result, we’ve been systematically scanning solutions which could solve the following question:</p><h4>How can you enable organizations to extract the most out of their data assets and stay in the clear?</h4><p>To do that we’ve been exploring the AI workflow stack from data preparation, pipeline, modeling, monitoring, and analyzed the limitations that exist today. The quality of output of AI algorithms depends on the quality of the data they were fed with and trained on, so it is crucial that the data preparation sample is an accurate representation of real-world life. Highly regulated industries are restricted from using personal identifiable information (PII), and with the rise of privacy laws (such as GDPR, CCPA), AI capabilities are inhibited.</p><h4><strong>Alternative techniques have certain limitations compared with synthetic data generation</strong></h4><p>Our team examined several technologies attempting to solve the issues described above. These include homomorphic encryption, anonymization, and synthetic data. While synthetic data can’t and won’t solve for every single privacy related use-case, it is likely to emerge as the <strong>most accurate and secure solution to address AI modeling and QA testing</strong> for customers trying to navigate the issue of data privacy and security.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fkkCxjMlCJTEvzHvP9QHiA.png" /><figcaption><strong>Source: TPY Capital analysis</strong></figcaption></figure><h4><strong>Datomize: a solution democratizing the use of private information and unlocking the potential for new business opportunities</strong></h4><p>Datomize generates high-quality synthetic data from existing datasets, while eliminating the risk of breaching sensitive personal information. One can think of it as a ‘digital twin’ of the data — on which you can work and run analyses, without exposing data you should not. Datomize provides its customers with an engine that takes their sensitive data and turns it into a synthetic replica, ready to be fed into AI/ML models.</p><p>Synthetic data is arguably the best solution today for use-cases where existing (real-world) data cannot be used or is not available to the required extent. Datomize’s platform <em>synthesizes </em>the existing dataset into a simulated, unidentifiable dataset for business applications that include development and training of AI models, QA testing, internal and external data collaboration.</p><figure><img alt="Source: Datomize" src="https://cdn-images-1.medium.com/max/607/1*KNuSV1go6IwB2V-Pp4ef5w.png" /><figcaption><strong><em>Source: Datomize</em></strong></figcaption></figure><h4><strong>The founding team balances a successful entrepreneurship track record with technological depth</strong></h4><p><a href="https://www.linkedin.com/in/avweiss/">Avi Weiss</a>, the co-founder and CEO of Datomize, impressed us with his experience, transparency, and passion for the problem Datomize is solving.</p><p>As we dug further, we’d met with the broader founding team, including <a href="https://www.linkedin.com/in/sigalshaked/">Dr. Sigal Shaked </a>(CTO) and <a href="https://www.linkedin.com/in/royyogev">Roy Yogev</a> (VP BD &amp; Marketing). We liked the breadth of the perspectives they brought to the table. Dr. Sigal Shaked has an extensive background in the field, both in academia and in industry. Roy Yogev brings experience working with Fortune 500 companies as well as channels like Wipro and other IT service providers, which are important for Datomize’s go-to-market strategy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*5gms4WciMY_9NtgxBbMihQ.jpeg" /><figcaption>from Left: Roy, Sigal and Avi</figcaption></figure><p>We’re glad to lead this funding round, with the participation of Datomize’s existing investors <a href="https://www.f2vc.com/">F2 Venture Capital</a> and are eager to see how this team revolutionizes how regulated industries collect, apply, and share data while protecting our privacy.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d8ceb20e1c4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/why-we-invested-in-datomize-d8ceb20e1c4">Why We Invested in Datomize</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in UNIPaaS (spoiler: it’s the next big thing in B2B payments)]]></title>
            <link>https://medium.com/tpy-capital/why-we-invested-in-unipaas-spoiler-its-the-next-big-thing-in-b2b-payments-7864f3bbcdbe?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/7864f3bbcdbe</guid>
            <category><![CDATA[payments]]></category>
            <category><![CDATA[venture-capital]]></category>
            <category><![CDATA[marketplaces]]></category>
            <category><![CDATA[investment]]></category>
            <category><![CDATA[fintech]]></category>
            <dc:creator><![CDATA[Guy Yamen]]></dc:creator>
            <pubDate>Thu, 11 Feb 2021 12:04:18 GMT</pubDate>
            <atom:updated>2021-02-11T12:08:47.814Z</atom:updated>
            <content:encoded><![CDATA[<p><em>Before anything else, if you’re a </em><strong><em>UK-based B2B digital platform </em></strong><em>and you’re looking for a comprehensive payments solution, </em><a href="https://portal.unipaas.com/signup"><em>sign up</em></a><em>. UNIPaaS’s portal is live!</em></p><p>And now, back to where we haven’t started off yet!</p><p><strong>Why we invested in UNIPaaS</strong></p><p>David Avgi came marching into our office in the late summer of 2020, alongside the team he’s been working with on his new venture.</p><p>By the time he had left, two hours after, we knew we wanted to do something with him.</p><p>Since then, we had gone much deeper into analyzing the B2B payments space. And coming out of this dive we realized we just HAD TO do something together.</p><p>This short piece will describe why we got so excited in the first place, and why this initial excitement translated into a more analytics-driven appreciation of what UNIPaaS is about to do.</p><p><strong>First meeting: September 2020</strong></p><p>We’re in our conference room, masks on our faces (a refreshing moment nevertheless; meeting in person! six months into COVID!). The UNIPaaS team members know each other very well. They’ve all worked together in Safecharge, the leading payments technology company which Avgi <a href="https://www.safecharge.com/press-releases/nuvei-completes-acquisition-of-safecharge-for-us-889-million-strengthening-its-global-payment-technology-and-reach/">sold to Nuvei</a> for ~$900m back in 2019. They start talking, and we realize that in between the team members, there’s more than fifty years of payments experience.</p><p>They paint a stark picture of what’s currently broken in the B2B marketplaces’ payments space.</p><p>They share that the growth of digital platforms, and marketplaces in particular, is exponential, not least because of the pandemic. They tell us that the streamlining, onboarding, and workflow of marketplaces are fundamentally different than those of traditional DTC or online platforms. They continue that existing offerings seem to fall short of the unique needs marketplace customers have.</p><p>David decided to name the new venture UNIPaaS, which stands for UNIversal, UNIfied, and UNIque Payments as a Service, and as the session progresses, the same question creeps into the mind of each of the TPY team members present: “why do I get the feeling that UNI could also stand for UNIcorn”?</p><p>Then David shares with us that he’d already invested a considerable sum to launch UNIPaaS, and that he intends to invest side-by-side the lead investor, in a meaningful way.</p><p>Let’s pause on this for a moment.</p><p>There is some lingering debate in the VC world of whether having a successful repeat entrepreneur invest in his/her own new startup is a positive thing. One argument ‘against’ refers to the risk of incurring somewhat of a bi-polar mindset by the CEO, who’s both entrepreneur and investor. Another involves theories around what a healthy percent stake for an entrepreneur should be.</p><p>Our perspective is that the positives far outweigh the negatives. By, quite literally, ‘putting the money where the mouth is’, David demonstrated his level of excitement from this venture, showed his commitment to its success, and enhanced the alignment of interests between his investors and himself. Simply put, we at TPY loved seeing it.</p><p><strong>From first meeting to closing: last quarter of 2020</strong></p><p>Over the course of the following few months we’d spoken to dozens of practitioners in the digital payments space and to many others in marketplaces. The feedback we received was overwhelmingly similar: that the pain point described by David and his team was real, and that UNIPaaS’s approach to solving it made a ton of sense.</p><p><em>So what’s so compelling in the solution?</em></p><p>Fundamentally, that it is indeed a <strong>Payments Platform that can be offered as a service.</strong></p><p>There’s no ‘deep tech’ in payments. No AI dark magic or a novel piece of hardware. UNIPaaS’s secret sauce is in the depth of its offering and its speed of execution, achievable only given the vast and intimate knowledge the team has of this space.</p><p>A year from kickoff, the team had developed a fully operational gateway engine, already processing thousands of transactions daily, with multi-channel (cards present; cards not present) connectivity to an acquirer (cards and APMs).</p><p>It launched a transparent billing system, offering clients real-time visibility into where the money is, and the exact fee and commission status. And it created a click-of-a-button API for pay-in and pay-out.</p><p>All written in-house, from scratch, tailored to this unique market’s needs, and based on best-in-class architecture. As a result, the platform is reducing time to payments for clients, and is built for scale, from day one.</p><p>By the end of the year, UNIPaaS will have developed a managed solution enabling marketplaces to seamlessly and effortlessly onboard sellers, vendors and contractors. It will accept payments on their behalf, enable splitting them, transfer funds, and do payouts. In short, marketplaces will have complete control over both buyer and seller experience while staying ahead of PCI /PSD2 strict regulation.</p><p>See a rough schema of what we believe UNIPaaS’s offering will look like by end of 2021:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Tnzibos7revubnbN7G1K2w.png" /></figure><p>And on top of all this, UNIPaaS obtained the UK’s regulator Payments Institution (PI) license, which enables UNIPaaS to support UK clients.</p><p>As investors, seeing the combined progress UNIPaaS has been making in product and operations in a regulatory environment has been jaw-dropping.</p><p><strong>Now let’s bring along the best of the best to this journey!</strong></p><p>Along the way, we introduced David to friends we thought were the absolute best B2B payments’ experts, and almost every single one of them had become an investor in this round. These included the fantastic Sarah Friar (CEO of NextDoor and formerly CFO of Square); the unstoppable Eric Sager (COO of Plaid); the fintech black belt guru Sam Hodges, CEO of Vouch Insurance and formerly President of Funding Circle; and the prolific product leader and angel investor Gokul Rajaram.</p><p>We’d also reached out to our friends at FJ Labs, the world’s leading fund in marketplaces investments, given their depth of knowledge of what future customers of UNIPaaS need, and now our partners in the UNIPaaS journey.</p><p>And when introduced David to some of our serial entrepreneurs of TPY’s portfolio companies, well, they all came back eager to participate. These included Talmon Marco (H2Pro; Juno; Viber) as well as Boaz Palgi and Erez Webman (Paramount Data; ScaleIO). It also included our friend Ofer Smadari (StackPulse; Luminate; not a TPY CEO…yet).</p><p>===</p><p>To conclude…the pain is there, the team is there, the alignment is there, the platform is there, and the clients are starting to come. And then there are the opportunities for UNIPaaS beyond payments, in expanding its offering to banking as a service and lending as a service, and the size of opportunity becomes truly huge (more on that in a separate future post).</p><p>We’re truly excited to see the evolution of this startup into what we hope will become a new industry standard.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/751/0*K5Kdkqq7ZgKS7r_w.jpeg" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7864f3bbcdbe" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/why-we-invested-in-unipaas-spoiler-its-the-next-big-thing-in-b2b-payments-7864f3bbcdbe">Why we invested in UNIPaaS (spoiler: it’s the next big thing in B2B payments)</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why TPY invested in a startup building a whole new database]]></title>
            <link>https://medium.com/tpy-capital/why-tpy-invested-in-a-startup-building-a-whole-new-database-2bbc27461c1c?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/2bbc27461c1c</guid>
            <category><![CDATA[vc]]></category>
            <category><![CDATA[rdbms]]></category>
            <category><![CDATA[nosql]]></category>
            <category><![CDATA[investment]]></category>
            <category><![CDATA[database]]></category>
            <dc:creator><![CDATA[Guy Yamen]]></dc:creator>
            <pubDate>Thu, 01 Oct 2020 14:03:43 GMT</pubDate>
            <atom:updated>2022-08-10T07:42:16.411Z</atom:updated>
            <content:encoded><![CDATA[<h3>Why we invested in a startup building an entirely new database</h3><p>A few months ago, we’d made an initial investment in Paramount Data (which had since changed its name to Regatta). Here are a few words on why we did.</p><p>We at TPY are fascinated by data. This draws our team to startups that generate data which did not exist before, or those which synthesize various types of data and make it available to generate actionable insights.</p><p>We also look at the usage of data to streamline and digitize organizational processes. And while it is undoubtedly less ’sexy’ at first sight, we do explore opportunities all the way down at the bottom of the data stack — where, among other peculiar creatures — PD resides.</p><p>At a high level, PD pertains to do something very tough: develop a whole new database. And developing a new database means exactly that: not an extension of an existing database, not a minor modification to an existing database, but a new database altogether.</p><p>Given this audacious task, we had to believe two things as an investment team:<br>a. that the existing solutions are not solving for what is a real and substantial pain <br>b. that this specific team can offer a viable solution to this pain</p><p><strong>Addressing a real, and BIG, pain point</strong></p><p>One can slice and dice the world of databases in many ways. A common one is the distinction between Relational databases and NoSQL ones;</p><p>Relational Database Management Systems (<strong>RDBMS</strong>s) support transaction consistency (ACID) and high performance but cannot represent various types of data and have limited scalability capabilities. Also known as SQL databases, RDBMS cannot run OLTP (transactional) and OLAP (analytics) in parallel because it would affect the database.</p><p><a href="https://en.wikipedia.org/wiki/Oracle_Database">Oracle</a>, <a href="https://www.mysql.com/">MySQL</a>, and <a href="https://azure.microsoft.com/en-us/services/sql-database/">Microsoft SQL</a> are all part of this family of databases.</p><p><strong>NoSQL</strong> databases provide more elasticity than relational databases. They are based on one of several data models supporting semi-structured or unstructured types of data, and at larger volumes. They do so not at the single database entity level, but rather on a cluster of servers, each acting as an independent sub-database.</p><p>Similar to their relational cousins, NoSQL databases are not very friendly to analytics heavy load. In addition, they have multiple architectural flaws, including lack of transaction consistency or ability to perform query operations across the sub-database servers’ boundaries. Both are crucial elements for enterprise-grade data infrastructures.</p><p>Among members of this group are <a href="https://www.mongodb.com/what-is-mongodb">MongoDB</a>, <a href="https://cassandra.apache.org/">Cassandra</a>, and <a href="https://www.couchbase.com/">Couchbase</a>.</p><p>See below a rough comparison of the two types of databases:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/592/1*1pyyOdoZ_29-RdspFffqHA.png" /><figcaption>TPY analysis</figcaption></figure><p>Then you have the newcomers or next-gen databases offering modern architecture databases This subset of offerings is known as <strong>NewSQL</strong>. The basic idea is to combine consistency (ACID) which RDBMS provides with the scalability and elasticity of NoSQL. We believe most NewSQL companies are still searching their way to delivering this promise, given that they were mostly built on top of existing (RDBMS or NoSQL) architecture.</p><p>Among the most popular NewSQL companies are <a href="https://cloud.google.com/spanner">Google Spanner</a>, <a href="https://mariadb.com/newsroom/press-releases/mariadb-acquires-clustrix-adding-distributed-database-technology/">MariaDB’s Clustrix</a>, and <a href="https://www.cockroachlabs.com/">Cockroach Labs</a>, and <a href="https://www.yugabyte.com/">YugabyteDB</a> (which are inspired by Google Spanner’s 2012 whitepaper).</p><p>So when we look at current day database deployment at the enterprise level, both RDBMS and NoSQL databases are typically being used in order to handle the unique limitations each of them imposes and given the clear tradeoff between the two.</p><p>And so, to support existing enterprise infrastructure with multiple types of data, developers are forced to deploy multiple “siloed” databases and use a mix of RDBMS and NoSQL.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/580/1*rf8_j7tCGUFQIJzmnPTQkA.jpeg" /></figure><p>As a result, it is impossible to get reports or run real-time analytics on data across siloed databases’, unless that data is duplicated into a data warehouse or a data lake. Such data duplication (sometimes called ETL) is operationally expensive and labor-intensive. When the enterprise needs to perform and support real-time data analytics, the inherent delay involved with ETL deems data lakes or data warehouses irrelevant.</p><p>===</p><p><strong>Enter PD, who is building a large-scale database that contains the best elements of both the relational and NoSQL databases and online analytics.</strong></p><p>We believe PD had already solved several technology challenges at the core of this industry, by developing a highly scalable, enterprise-grade, high performance, elastic database that offers:</p><ul><li>Grow &amp; shrink the database cluster elastically, ubiquitously to the user or application</li><li>Distributed ACID while maintaining high performance</li><li>Distributed compute (e.g. distributed JOIN)</li><li>Co-existence of OLTP and OLAP</li><li>Unified rich data model</li></ul><p><em>In a nutshell, PD’s software-only database brings together the benefits of relational databases, NoSQL databases, and data analytics/lakes/warehouses in a single cluster.</em></p><p>The initial sectors PD is tackling are financial services, government, web application enterprises, and pharmaceuticals. As the world is becoming more real-time, organizations are interested in faster application-time-to-market, higher performance applications, and more real-time insights into both hot and historical data.</p><p>As most industries today use large amounts of historical data they are willing to pay for solutions that can fundamentally improve their data architecture, simplify development, significantly expedite/shorten application-time-to-market, and enable unprecedented analytics capabilities and time-to-data-insights.</p><p><strong>Team. Team. Team.</strong></p><p>We’d met two of the three co-founders at a bar.</p><p>No, this is not the start of a joke but the surprising reality that good things can actually come out VC industry events!</p><p>What had struck us — beyond realizing that these two can carry their drinks — were a few elements which became only more obvious as we got to know the team better in the following months:</p><p>a. The founders had worked together in past startups, including the latest one, <a href="https://techcrunch.com/2013/06/19/emc-acquires-israeli-storage-startup-scaleio-for-200m-300m-to-compete-better-with-the-cloud-kings/">ScaleIO which they’d sold to EMC</a> a few years back. It maybe goes without saying, but there’s something very compelling</p><p>b. The founding team started working on PD January 2019, alongside five other team members, who had joined in the summer of 2019, and all have been working on both design and coding, with product development progressing according to plan. Our estimate is that no less than 55 months of cumulative work had been put into venture prior to the initial investment.</p><p>c. We loved the fact that from day one the team was seeking to identify commercial applications, rather than building the startup at the darkness of stealth, waiting for the design partnership to unveil itself after years of development.</p><p>d. As we got to know each other better, we were impressed by the team’s hard-working attitude; the practicality they showed in the negotiations with us; and the down to earth culture.</p><p>And when we say down to earth we mean it — PD is based out of a small village about an hour north of Tel Aviv, just next to a cowshed!</p><p>TPY is excited to partner with <a href="https://www.linkedin.com/in/palgi/">Boaz Palgi</a>, <a href="https://www.linkedin.com/in/erez-webman-20a7382/">Erez Webman</a>, <a href="https://www.linkedin.com/in/eran-borovik-41a5541/?originalSubdomain=il">Eran Borovik</a>, and the rest of the PD team in the journey to offer a whole new breed of a database.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2bbc27461c1c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/why-tpy-invested-in-a-startup-building-a-whole-new-database-2bbc27461c1c">Why TPY invested in a startup building a whole new database</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why we invested in XTEND, and why we doubled down just as COVID-19 hit us]]></title>
            <link>https://medium.com/tpy-capital/why-we-invested-in-xtend-and-why-we-doubled-down-just-as-covid-19-hit-us-ec806d6351b0?source=rss----9eff3e2c8cf2---4</link>
            <guid isPermaLink="false">https://medium.com/p/ec806d6351b0</guid>
            <category><![CDATA[drones]]></category>
            <category><![CDATA[defense]]></category>
            <category><![CDATA[startup-nation]]></category>
            <category><![CDATA[extended-reality]]></category>
            <category><![CDATA[vc]]></category>
            <dc:creator><![CDATA[Guy Yamen]]></dc:creator>
            <pubDate>Thu, 06 Aug 2020 09:18:48 GMT</pubDate>
            <atom:updated>2020-11-23T13:43:51.158Z</atom:updated>
            <content:encoded><![CDATA[<p><em>”Are you doing defense?!” “TPY is into drones now, ha?”</em></p><p>Well, the answer is a bit more complex…</p><p>Just prior to the COVID outbreak we made an initial investment in <a href="https://www.xtend.me/">XTEND</a> (t<a href="https://vimeo.com/444916137/3a27761ead">his video</a> gives a nice overview of its vision).</p><p>To those asking us what’s the rationale behind our investment, I will try to synthesize the very basics of our thesis.</p><ol><li><strong>The beauty of a robust tech platform supporting multiple potential use-cases</strong></li></ol><p>In general, the platform/use-case play in startups is tricky.</p><p>At one extreme, a pure horizontal technology platform play can be a formidable force. This is because almost every single dollar deployed by the startup, goes to development, which can help create an IP-based defensive moat around it. See as one example, <a href="https://techcrunch.com/2014/01/26/google-deepmind/">DeepMind, a horizontal AI platform which ended up being acquired by Google</a> for (reportedly) more than $500m, four years after inception.</p><p>Ultimately though, it’s a binary play: you either make it very big, as a de-facto standard in your industry, and become attractive to acquirers who want to plug in your ‘engine’ into their commercial workflows, or you’re left stuck with that shiny engine you’ve built, realizing it has no commercial relevance or application.</p><p>At the other extreme, there is the singular use-case startup, going after a niche vertical market.</p><p>The advantages of such play are straightforward: such startup can demonstrate product-market-fit very fast, build traction, and create a defensible commercial moat, if not necessarily a deep tech one. Most earlier stage startups we come across fall into this category.</p><p>What we saw in XTEND is an opportunity for a compelling <em>hybrid model</em>. XTEND’s underlying technology enables ANY user to fly a drone accurately and seamlessly, with minimal training.</p><p>This means accuracy, zero-learning time, transitioning between indoor &amp; outdoor, and translation of users’ hand gestures of where to fly the drone in a physical environment.</p><p>The technology supports extremely low latency (depending on the distance) and high-quality video transmission, which allows for deep immersion and spatial (visual) awareness.</p><p>This platform, the XTEND team had realized, lends itself very nicely to the world of <strong>defense</strong>. Allowing military forces to launch a drone to survey indoor parameters, as one example, or to intercept low altitude flying aerial threats, without putting its personnel in harm’s way.</p><p>We thought the way XTEND had tackled this was compelling: if the market is there, pursue it, but do so in a contained manner.</p><p>Prior to our investment, XTEND created a fully owned subsidiary focusing exclusively on defense. This translated to a siloed commercial unit, sitting on top of a broader technology chassis. This chassis, we believe can be <em>extended</em> into multiple plays. Reaching all the way from fields adjacent to defense (e.g., HLS — Homeland Security) to very remote ones (e.g., gaming).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tcroUijS84_wb7bSxBY2sw.jpeg" /></figure><p><strong>2. Topnotch execution</strong></p><p>This parameter is not an easy one to objectively pinpoint to, but from the moment we started interacting with the team we observed the kind of qualities we really like: a persistent, go get’em kind of attitude, which is single-mindedly focused on deliverables.</p><p>There’s nothing standard in the way XTEND operates, but somehow, in between the humming of drones and the chirping of VR (virtual reality) equipment, one can detect discipline in execution which could be traced back to the last startup the dynamic-duo of siblings behind XTEND founded, and which they <a href="https://en.globes.co.il/en/article-intel-acquires-israeli-co-replay-technologies-for-175m-1001108929">had sold to intel</a>.</p><p>==</p><p>And so, in Q2 — a mere few weeks after our initial investment, when the world was turning upside down and the company decided to raise additional capital, TPY had doubled down on its investment in XTEND.</p><p><strong>How come?</strong></p><p>For one, we had a broader view into the execution the team is capable of, which gave us a good perspective that our pre-investment thesis was on spot on this front.</p><p>But more holistically, we see COVID-19 unleashing two trends which are very positive for XTEND:</p><ol><li><strong>Remoteness</strong>: in a nutshell, people are doing less <em>on prem</em>, and more <em>off prem</em>. This will not be going away even as the virus does, and it will be extending itself from defense to areas such as predictive maintenance in manufacturing or even, imagine that, tourism (if you can’t go to Rome, why not ‘be’ there by sitting in your living-room-sofa, using your VR glasses and flying an XTEND drone which is by the Colosseum).</li><li><strong>Security</strong>: it’s an unpleasant reality of the crisis we’re going through, but we’ll need more first responders with far less federal or state budgets available to deploy them. XTEND’s technology can help bridge this supply-demand gap. Imagine an ability to provide instantaneous first response &amp; security services to the wide population, irrespective of physical distance to the one in need (for example a police officer can project herself into a drone to a burglary site).</li></ol><p>Is this a whole new way for us to experience the world? a bold version of what human-machine-interface looks like? We do hope so.</p><p>And as we’re kicking off this partnership with XTEND, we wish them, and us, good luck — after all, they’re the reason we’re present in telepresence!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec806d6351b0" width="1" height="1" alt=""><hr><p><a href="https://medium.com/tpy-capital/why-we-invested-in-xtend-and-why-we-doubled-down-just-as-covid-19-hit-us-ec806d6351b0">Why we invested in XTEND, and why we doubled down just as COVID-19 hit us</a> was originally published in <a href="https://medium.com/tpy-capital">TPY Capital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>