Articles by Sundeep
-
Advancing medicine through real world data
Advancing medicine through real world data
Anyone impacted by chronic/rare diseases knows the frustration of dealing with dispersed healthcare records. Dozens of…
11 Comments -
Komodo Health: The Anti-Unicorn
Komodo Health: The Anti-Unicorn
Today, Komodo Health announced it's Series C raise, led by A16Z and Oak. Komodo's mission is to reduce the global…
8 Comments -
Introducing InCountry: Because Your Data Shouldn’t Have a Passport
Introducing InCountry: Because Your Data Shouldn’t Have a Passport
In recent years, GDPR has become a household acronym and major global data privacy mishaps have increased in frequency…
1 Comment
Activity
7K followers
-
Sundeep Peechu reposted thisSundeep Peechu reposted thisThe loudest story about AI is a lonely one. One person with an army of chatbots. Other humans are friction. That gets the future wrong. The best things aren’t built alone. In a moment of change, we want to remind the world (and ourselves) what Notion stands for: — Think Together
-
Sundeep Peechu shared thisSundeep Peechu shared thisHow do you know which AI model to trust? That's not a rhetorical question. It's one of the most important infrastructure challenges in AI right now. Anastasios Angelopoulos and Wei-Lin Chiang built Arena to answer it. What started as a Berkeley PhD research project to rank chatbots by human preference has become the backbone of how the world evaluates and advances AI. Frontier labs like OpenAI test models on the platform before public release. Researchers, developers, and builders around the world use it to understand what AI can actually do, and not just what benchmarks say it can do. The insight at the core of Arena: traditional benchmarks are easy to game. You can build for benchmark scores while ignoring real-world performance. Arena uses crowdsourced human preference data at scale to produce something harder to fake - a signal that actually reflects how models perform when people use them. We at Felicis have been believers since the seed round, and proudly led the Series A. Read the full story of how Anastasios and Wei-Lin built AI's most trusted evaluation platform — and why that matters more as the models get more powerful. ordnl.link/ywOwWNI
-
Sundeep Peechu reposted thisSundeep Peechu reposted thisToday, we launched a fresh new look for Felicis built on the core belief that has guided us from the beginning: Fortune favors the brave. Founding a company is an act of courage. We are the true believers in founders who have the imagination, bravery, and discipline to defy the odds and ~build something extraordinary~. We focused on clarifying what makes us different in a sea of VC sameness. Lucky for us, Felicis is a pretty unique firm! We are relentlessly founder first. Unabashedly optimistic. Full of imagination. We move fast and early. We prioritize first principles over consensus. We build trust into every touchpoint. We believe you engineer your own luck. Our track record backs it up: - 0 board votes against our founders. We vote with founders. Always. - 1% on top of every first check goes to founder support. We give founders what they need to become unbreakable. - 24 hours from meeting to term sheet. We move at the speed of founders. - 80% of the founders we back are referrals from our portfolio. Founders know what founders want. - 94% of our investments are Seed or Series A. We invest before it’s obvious, when founders are just beginning to build. - 83% of the time, Felicis leads or co-leads. We back what we believe in, early and without hesitation. - 40+ home countries of portfolio founders. Greatness can come from anywhere. Our unique approach has helped Felicis-backed founders build generational companies. Since our founding in 2006, we've backed 50+ companies now valued at $1B+, 125+ portfolio companies have been acquired or gone public, and our portfolio has captured $285B+ in market value. We are among the top 8 VC firms with a track record of backing mega-unicorns ($5B+ exits) at Seed, A, or B in the last decade. And 11% of Felicis backed-companies have become unicorns after we invest – compared to a 1% industry average. I’m proud of the team who brought our brand to life, and so grateful to the founders, LPs, and Felicis community who inspire it every day. Take a look. I’d love to hear what you think! 🔗 felicis.com Brand Strategy, Creative, and Design: Play SF 🧠 🎨 Production Design: KBD 💪 Content: Colette Keane ✍ Team photography: Joseph Seif 🐐 📸
-
Sundeep Peechu reposted thisSundeep Peechu reposted thisWe started PermitFlow because permitting is one of the most overlooked barriers to housing affordability in America. Today's White House Executive Order addresses this directly, and I’m honored our team could contribute to this work. Section 4 of the Executive Order, "Removing Regulatory Barriers to Affordable Home Construction," reflects some of our recommendations: "Streamlining permitting processes for housing developments by, for example, capping permitting timelines and fees; allowing by-right development for single-family homes; limiting retroactive application of new or changed building codes." Months of effort across our team went into this: → Deep analysis of permitting and regulatory trends across our proprietary datasets → Evaluation and recommendation of solutions to improve affordability → Multiple onsite White House working sessions with talented administration policymakers This Executive Order is an exciting start. We're committed to continuing this work and grateful for the chance to contribute. Read more about the Executive Order here: https://lnkd.in/gf9C2jxyRemoving Regulatory Barriers to Affordable Home ConstructionRemoving Regulatory Barriers to Affordable Home Construction
-
Sundeep Peechu shared thisI asked Brendan Foody from Mercor a simple question on stage at #UpfrontSummit: why hasn't enterprise AI had its ChatGPT moment yet? His answer: the data doesn't exist yet. And humans have to create it. The internet trained models to search. GitHub trained them to code. But no one has written down how a consultant, lawyer, or banker actually thinks. That context lives in people's heads, and unlocking it is the defining challenge of the AI era. Brendan's bet: the future of work is humans training agents to handle the redundant stuff, so they can focus on the bigger unsolved problems: "There's no shortage of problems we need to solve. We need to cure cancer, solve climate change, go to Mars." And, according to Brendan, it will mean people and computers working together to get there. "The irony is that the future of AI is actually very human." I'll share the video of the full conversation when it's up. Thanks to Mark Suster and the Upfront Ventures team for including us in such a fantastic event.
-
Sundeep Peechu reposted thisSundeep Peechu reposted thisSome incredible stats from Cliff Obrecht on Canva in 2025: 265M MAU (31M paying users) $4B ARR B2B (25 seats or more) now at $500M ARR (Growing 100% yoy) AI feature that lets users build apps and websites hit 10M MAUs… 🚀🔥🤩 https://lnkd.in/gH6FpYXFCanva gets to $4B in revenue as LLM referral traffic rises | TechCrunchCanva gets to $4B in revenue as LLM referral traffic rises | TechCrunch
-
Sundeep Peechu reposted thisI’m thrilled to step into my new role as General Partner at Felicis. To the founders who’ve trusted me so far: thank you. I don’t take it lightly, and I’m committed to earning that trust every day. And to future founders: if you have an insight that feels obvious to you and invisible to everyone else, let's meet!Sundeep Peechu reposted thisWe’re excited to share that James Detweiler has been promoted to General Partner at Felicis! 🧡 James is thesis-driven, founder-first, and unafraid to challenge convention. He is unique in his ability to see around corners and believe in transformative technology before consensus. He has an exceptional ability to identify and earn the trust of founders and a strong investing track record across AI, robotics, vertical software, and defense. An example: His early conviction around robotics foundation models led to Felicis investing in the Skild AI Series A. James has helped bring companies like Mercor, Paraform, DeepJudge, Ricursive Intelligence, and Operand into the Felicis family. Beyond the results, James shows up with infectious energy, generosity, and optimism. He lifts the entire team and goes all in for founders every step of the way. We’re proud to recognize his impact and excited for what’s ahead! ordnl.link/yFDDZZE
-
Sundeep Peechu reposted thisThrilled that James Detweiler has been promoted to General Partner at Felicis !! James helps all of us expand our imagination of what's possible. He’s thesis-driven, unafraid to challenge convention, and brings fresh perspectives that push us to see around corners. His ability to earn founder trust is exceptional, and his track record speaks for itself: 25+ investments across AI, robotics, vertical software, and defense, including Mercor , Paraform , Skild AI , DeepJudge , Ricursive Intelligence , and Operand . I still remember the first time James walked me through his robotics thesis. It completely reframed how I thought about the opportunity. He brought @Skild AI to Felicis at Series A based on an early belief that robotics foundation models would create a “ChatGPT moment” for robotics. But what really sets James apart is how he shows up. He’ll travel anywhere to meet founders, go deep on the hardest problems, and give 100% to the teams he backs. He lifts everyone around him with his energy, generosity, and optimism. If you’re a founder working on foundational AI or ‘root node’ problems, James is someone you want in your corner. Huge congrats, James !!Sundeep Peechu reposted thisWe’re excited to share that James Detweiler has been promoted to General Partner at Felicis! 🧡 James is thesis-driven, founder-first, and unafraid to challenge convention. He is unique in his ability to see around corners and believe in transformative technology before consensus. He has an exceptional ability to identify and earn the trust of founders and a strong investing track record across AI, robotics, vertical software, and defense. An example: His early conviction around robotics foundation models led to Felicis investing in the Skild AI Series A. James has helped bring companies like Mercor, Paraform, DeepJudge, Ricursive Intelligence, and Operand into the Felicis family. Beyond the results, James shows up with infectious energy, generosity, and optimism. He lifts the entire team and goes all in for founders every step of the way. We’re proud to recognize his impact and excited for what’s ahead! ordnl.link/yFDDZZE
-
Sundeep Peechu reposted thisSundeep Peechu reposted thisEvery year, the radiology world descends on Chicago for RSNA to share research and ideas. But here’s the real question: which of these ideas will actually change how people understand their health before they get sick? Tens of thousands of radiologists. Hundreds (or, thousands) of abstracts. An overwhelming amount of genuinely exciting innovation in imaging, AI, and medicine. This year, Prenuvo presented several data-driven studies—but one result cut through the noise in a big way. Our abstract was the #1 most covered scientific press release at RSNA 2025. - Audience impressions: 1.75 billion - Media placements: 1,493 The headline finding: 1) Higher muscle volume is associated with a younger brain. 2) Higher visceral fat (relative to muscle) is associated with an older brain. Using whole-body MRI and brain MRI across 1,164 healthy adults, the team showed: - More muscle → lower chronological age and lower predicted brain age - More visceral fat per unit of muscle → higher chronological age and higher brain age - Subcutaneous fat? Not meaningfully associated In plain English: - Body composition shows up in the brain. What makes this powerful isn’t just the correlation—it’s how it was measured: - Direct MRI quantification of muscle and fat (not proxies) - AI-derived brain age models trained on thousands of scans - A whole-body view instead of isolated organ silos This is exactly why proactive imaging matters. Not to just “hunt for disease,” but to surface early signals, establish baselines, and connect systems that traditional care treats separately. Huge credit to Cyrus Raji, Sam Hashemi and the broader Prenuvo research team for pushing this work forward—and for showing how imaging can help us understand aging before symptoms show up. More to come.
-
Sundeep Peechu liked thisSundeep Peechu liked thisFor marital bliss, choose Prenuvo. 😂 🤣 😂 Proving that not all life-changing impact needs to come from early detection of the big “C.” Sometimes, the value shows up in quieter, more human ways. Here’s a message from a doctor who referred a patient to our clinic in London: "I sent a patient to have a full body MRI at Prenuvo. Not your typical (dramatic) cancer diagnosis. A 'simple' deviated septum found. Which allowed me to double down on his sleep (he wasn't keen on it initially). We did a sleep study and found he was not sleeping well. He had the septoplasty done and it made a huge improvement in his sleep. He's a different man now!" I’m told his wife—who had to put up with him snoring for years—loves us.
-
Sundeep Peechu liked thisSundeep Peechu liked thisSupabase just hit 100K stars on GitHub. GitHub stars are a vanity metric, but I love open source so it's something that matters to me. Thank you to all the contributors, maintainers, and partners. we didn't do it alone - we're standing on the shoulders of elephants Blog post: https://lnkd.in/eiFKNacv
-
Sundeep Peechu liked thisSundeep Peechu liked thisCongratulations to the Felicis portfolio companies recognized on the Wing Venture Capital on the 2026 Enterprise Tech 30 👏 This list includes the most promising builders shaping the future of enterprise; those at the forefront of the AI and agentic supercycle redefining how businesses operate. From design to developer tools to healthcare, these teams are pushing the boundaries of what’s possible with AI. Congrats to Canva, n8n, Supabase, and Assort Health on this well-deserved recognition. ordnl.link/UKykWaQ
-
Sundeep Peechu liked thisSundeep Peechu liked this"AI security will be bigger than cloud security." I heard some version of this from nearly every CISO I talked to at #RSA2026 last week — and I talked to a lot of them! (Worth noting: we at Felicis called this two years ago: Our 2024 research surveyed 40+ CISOs and came back with the same conclusion. The difference is that back then, it felt like a prediction. Now it feels like an emergency.) The vibe at RSA was less "here's our strategy" and more controlled anxiety. CISOs are caught in a genuine paradox: they're being told to adopt agents to stay competitive, while simultaneously having no real visibility into the AI and agents already running across their enterprise. That's the security void. The evolution happened fast and mostly sideways. It started with giving employees secure access to ChatGPT. Then it was firewalling copilots from sensitive internal data. Now every CISO I spoke with has woken up to the same uncomfortable truth: they don't know what's running, which means they can't assess the risk. The same challenges kept coming up: – "Everyone is struggling to secure AI while simultaneously adopting agents." – "We have to use agents or risk falling behind or losing jobs." – "There is no security playbook yet." That last one matters. The data backs it up: according to Gartner, 74% of Security/IT leaders agree that agents represent a new attack vector, and 84% say they need additional technical controls to govern and secure them. Dark Reading found that 48% of respondents expect agents to be the top attack vector by end of 2026. And Bessemer has called securing agents "the defining cybersecurity challenge of 2026." Anecdotally, from my own conversations at RSA: virtually all of the CISOs I spoke with are prioritizing spend here. The urgency is not theoretical. What's missing is action. The gap between how critical this problem is and how little has actually been purchased to solve it is enormous. Most CISOs know agent security is overdue. Very few have something in place. The space is moving too fast, and the category is still taking shape. That's exactly the kind of market condition that creates a "new Wiz." Right now, I think that opportunity is wide open and companies like Felicis-backed Runlayer are in a position to claim. The CISOs I spoke with are looking for something that solves the holy trinity: security across AI, identity, and data, in a world increasingly run by agents. The questions I'd be pressure-testing if I were a CISO right now: – Visibility: What AI and agents do I have across my enterprise? – Data & Identity: What do they have access to? – Governance: What rules govern their behavior? – Observability: Are we monitoring what they're doing? – Security: Are we detecting and remediating when something goes wrong? ordnl.link/mVfoN1W
-
Sundeep Peechu liked thisSundeep Peechu liked thisCompany building is usually about looking forward. So when I recently got the chance to guest lecture at Harvard College and Stanford University Graduate School of Business, it was a nice opportunity to reflect in the same halls where I sat as a student. I got to share stories from the early days of PermitFlow. Speaking to project managers after negotiating past reception with Krispy Kreme donuts. Talking to contractors about their permitting nightmares on the sidewalk outside the San Francisco Department of Building Inspection. That unglamorous, customer-centric street work shaped the company. Not by generating revenue - those weren't great sales tactics. But by deepening our empathy and connection with our customers. As PermitFlow scales toward $100mm+ ARR, those insights now flow from our incredible sales and customer success teams. But that scrappy, hands-on connection to builders remains our North Star.
-
Sundeep Peechu liked thisSundeep Peechu liked thisThe most technically gifted students aren't asking, "Which company should I join?" They're asking, "What should I build?" This past weekend, Felicis VC Fellows David Chung and Aarush Agarwal (multi-time winners across CalHacks, HackCMU, HackHarvard) organized and ran the first-ever Felicis x Carnegie Mellon University hackathon—bringing together 50+ highly technical students for an 8-hour build sprint. We kicked it off with a Q&A with Ananye Agarwal (founding engineer at Skild AI and CMU PhD). By the numbers: • 114 applications → we selected for top technical signal • 53 participants → 20 projects built and demoed • Talent included: Anthropic AI Safety Fellows, ML PhDs, incoming engineers at Anduril, Databricks, Snowflake, and researchers from CMU + MIT One thing that stood out: CMU students have real curiosity around building and joining something from the ground up. Huge credit to David and Aarush for building this from scratch. We're excited about CMU and more broadly about the next generation of builders who are deeply technical, curious about startups, and just starting to explore what they could build. We see what you see and are excited to dream with you!
-
Sundeep Peechu liked thisSundeep Peechu liked thisMost venture capitalists will tell you that consumer tech investing is hard. On the surface, they're right: consumer preferences are fickle, distribution is expensive, big platforms will copy anything that works, retention is brutal, monetization is uncertain. These are all true, but they name the symptoms, not the underlying reasons. Succeeding in consumer is difficult because it’s extremely rare to find a company that checks all the boxes. I’ve only done one consumer angel investment in the past 8 years because the perfect setup is hard to find. Wrote up my detailed thoughts here. Let me know what you think.
-
Sundeep Peechu liked thisSundeep Peechu liked thisour dear friend Amy Hillyard is missing. hoping some of you have seen the many posts + stories circulating the past 3 days, and was hoping that the search would have ended by now; but it’s continuing on, so we are hoping for more awareness throughout the bay area and beyond...if you’ve been part of anything I’ve been helping to build over the past two years, there’s a good chance Amy touched it ~ quietly. generously. pro bono. from flying to DC for #climatefoodhope and attending last year’s Foreign Cinema fundraiser to support our work with Alice Waters, Edible Schoolyard Project + School Supported Agriculture (all of which she believes in deeply!) ~ to shaping conversations about legacy and impact. She has also been a big supporter of She Can Global, Clio’s Bookstore Bar in Oakland, and Minerva University salons. she creates space for others, literally, with her husband Chris at Farley’s Cafe in Oakland and SF, where we gathered for a Galentine’s salon #135 last February 13th rooted in love, connection and care. this is someone who shows up for people. now we need to show up for her, in any way we can… if you have any information at all, please reach out immediately. 🙏🏽
Other similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content