About
📧 james@felicis.com
📞 p₁₆ × p₄₅₁₂₇₃₇
📞 p₁₆ × p₄₅₁₂₇₃₇
Activity
16K followers
-
James Detweiler shared thisJames Detweiler shared thisNearly every advanced technology, from EVs to the precision weapons the US depends on, runs on rare earth magnets. Almost none of that supply chain runs outside China, and beginning in 2027 US defense contractors will have to certify that no stage of theirs passed through China. Solcoa Industries is building the missing piece. From its warehouse in Alameda, it makes magnet-grade rare earth metal from oxides and from scrap, at lower cost than the conventional route and without greenhouse gases. It's the step the US has spent years and more than a billion dollars trying to rebuild at home, and it sits at the center of our global resilience focus. We're proud to keep backing Hooman Reza Nezhad and Artem Iurkovskyi by investing in Solcoa's Seed. James Detweiler tells the story here: felicis.link/dP9AmDd
-
James Detweiler shared thisIn a warehouse in Alameda, I watched Hooman Reza Nezhad hold a crowbar up to a magnet that used to power an MRI machine. The magnet ripped the bar out of his hands, pinning it with a loud crack. That magnet would soon be melted down to recover its neodymium and praseodymium, metals at the heart of every EV motor, drone, wind turbine, and precision weapon America builds. High-performance magnets need them, and almost no one outside China can turn rare earths into metal pure enough to make them. Solcoa Industries can. Hooman and his co-founder Artem Iurkovskyi have reinvented the chemistry that turns rare earth and scrap into magnet-grade metal. They do it at lower cost than conventional methods, without emitting greenhouse gases. The US hasn't commercialized a new way to make a primary metal since the 1940s. Solcoa is doing it today. A year ago they made this metal by hand, a few grams at a time. Today the line runs at 10 tonnes a year and their first commercial plant is built for 500. America wins by inventing. Felicis is proud to invest in Solcoa's Seed.
-
James Detweiler shared thisI flew to Pittsburgh to meet Deepak Pathak and Abhinav Gupta before Skild AI Series A, and it was immediately clear this was a special team tackling one of the hardest problems in AI: getting robots to learn like humans do. Today's news is why we at Felicis have been true believers in Skild since the beginning. S1 learned to play soccer entirely through self-play, competing against its own past versions across 140+ years of simulated play. It discovered dribbling, shielding, tackling, recovery, and even early teamwork with zero human demonstrations. This is the same recursive self-improvement that took AlphaGo from novice to superhuman, now arriving in the physical world. Soccer is just the first proof point on the way to something much bigger. Incredible progress for robotics, and incredible execution by the whole Skild AI team. Congrats! 🤖 ⚽ cc: Aydin SenkutJames Detweiler shared thisWe trained a robot to play football. How? Self-play for 140 years in a virtual World Cup. It was born in a physics simulation. For 140 years, our model played against increasingly tough opponents, each one a previous version of itself. Once it was ready, we challenged it to a match. ⚽ AlphaGo taught us that self-play reinforcement learning can lead to emergent, superhuman behavior. We're bringing it to the physical world. Blog post: https://lnkd.in/geTXzTMe
-
James Detweiler shared thisPhysical AI has a demo disease. The cure is deployment. Skild's execution is of the highest order: 60+ paying customers and $100M in ARR, just ten months. And the vast majority of that revenue is already coming from manipulation-related use cases. The team has also shown why deployment isn’t downstream of robotics research; it is itself a research frontier. Working in real-world environments revealed that customers need accuracy, speed, adaptability, and the ability to learn new tasks quickly. These insights directly shaped S1 and its approach to in-context learning. Congrats to the Skild team - this is the tip of the iceberg! We at Felicis are proud to back Deepak Pathak , Abhinav Gupta and the entire Skild AI team.James Detweiler shared thisWe just hit 100M ARR, 10 months after our first deployment. One of the fastest growing physical companies in human history. Deeply grateful to our team and partners for making this achievement possible. We're just getting started. More details: https://lnkd.in/ghRckdyr
-
James Detweiler shared thisTwo years ago, I flew to Pittsburgh to meet Deepak Pathak and Abhinav Gupta at CMU. I left convinced robotics foundation models would eventually yield a ChatGPT moment for the physical world. I didn't expect it to arrive this fast. Give S1, Skild AI's latest model, one video demonstration and it can execute an unseen task for up to ten minutes, without task-specific fine-tuning or post-training. Skild found that matching that performance with conventional post-training took roughly 380 examples. We at Felicis backed Deepak, Abhinav, and the team at the Series A on exactly this idea. Watching them execute has been an absolute joy.James Detweiler shared thisIntroducing S1, our general robot model that learns from just one example. S1 can carry out 10 minute long tasks it was never trained on, simply by prompting it with a single human demo without fine-tuning. How does it work? Much like a language model, S1 does not update its weights to learn a new task. Instead, it receives a demonstration of the task in its context window, and uses that to decide what to do next. It can be taught extremely long-horizon tasks. By composing manipulation primitives learned in pre-training, it can make a new style of coffee, pot a plant, fry pancakes, and more. The first time S1 flipped a pancake, we assumed pancake flipping must have been in its pre-training data. We searched millions of hours of pre-training data — and found nothing. S1 inferred the out-of-distribution task from one demonstration. Compared to conventional VLAs, S1 is a step change improvement. On unseen tasks after 100K hours of pre-training, language-prompted VLAs reach 9% accuracy. Our in-context learner reaches 66%. S1 does not rigidly replay the demonstration. A model of its size can withstand perturbations, recover from its own errors, and at times execute with more precision than the demonstrator. We're excited to deploy S1 with our limited industrial partners today, rolling out to all customers over the next few months. S1 is the first light on a new path towards general intelligence — grounded in the physical world. Technical blog: https://skild.ai/blogs/s1
-
James Detweiler shared thisI've spent a lot of time with talented founders. Far fewer who feel generational. Mercor founders Brendan Foody, Adarsh Hiremath, and Surya Midha are in that rare category. Mercor has gone from a dorm-room idea to a $2B run-rate business in just three years. The growth is extraordinary, but even more impressive are the founders behind it. When Brendan, Adarsh, and Surya met the founding team at xAI in 2023, "human data" wasn't even a category yet. But they saw where AI was headed: the next leap wouldn't come from more compute alone. It would come from expert human intelligence. So they started building the network the AI industry would eventually need. Today, Mercor has built a network of more than 5 million experts and pays out over $4 million every day to the people whose knowledge helps generate, validate, and evaluate the data that makes frontier models smarter. That may prove to be one of the most important infrastructure layers of the AI era. We backed Mercor because we believed the market for human expertise was fundamentally broken, and that fixing it would become foundational to how AI gets built. Just as important, we believed in them as founders. They move with incredible speed, obsess over customers, and meet challenges with a level of transparency and maturity that's rare at any age. We’ve been true believers in Mercor since the Series B, and we're even more bullish today.James Detweiler shared thisThere's no website that captures how a radiologist interprets a scan, how a banker sizes a deal, or how a lawyer builds an argument. That knowledge lives in people's heads. Mercor’s mission is to organize human intelligence to power the AI economy. Brendan Foody, Adarsh Hiremath, and Surya Midha built the network that finds the world's best experts, puts them to work training AI, and pays them for it, to the tune of $4M+ a day. From chess masters, radiologists, chefs, mathematicians … name a niche and they have it covered. Which is why nearly every top AI lab is a customer. We at Felicis are true believers in Mercor , the fastest startup to ever reach a $500 million revenue run rate. In March 2026, they reached $1 billion. And by June, Mercor had reached $2 billion in annualized revenue run-rate. Read the full story from Matt Quinn on how they went from college sophomores who started in a dorm room in 2023, to building core infrastructure for the AI economy today: 🔗 felicis.link/qPQCnhM cc: Sundeep Peechu James Detweiler
-
James Detweiler reposted thisJames Detweiler reposted thisParaform crossed $100M in annualized revenue just two years after we made our first dollar - and we're accelerating quickly, 2x’ing revenue in the last 6 months alone. To our customers - thank you for trusting us with your most valuable asset: your people. We know every hire shapes the future of your company, and it's a responsibility we don’t take lightly. We’re honored to play a part in helping you build your team. We'll continue to push the boundaries of what's possible in hiring. When we started Paraform, we expected our earliest customers to be startups and technology companies. What surprised us was how quickly organizations across entirely different industries - from Fortune 500 companies to AmLaw 100 law firms - embraced a new way of hiring. It reinforced something we now believe deeply: building exceptional teams is a universal challenge. To our recruiters - Paraform would not exist without you. You are the bedrock of what we have built. No software, algorithm, or AI can replace the trust and relationships you create with candidates. Some of the most meaningful moments building this company are hearing how Paraform has helped recruiters build businesses of their own, spend more time with their families, pay off debt, pay for their weddings, buy homes, and reach milestones they never thought were possible. We started Paraform to solve a broken hiring process. Along the way, we realized we also had the opportunity to help rebuild a recruiting industry that hasn’t always worked for recruiters. Our goal is to put more power back in your hands, giving you the opportunities and infrastructure to build a real business by focusing on what you do best: making great placements. We will keep building for you and investing in your success. To our team - this milestone belongs to you. You work incredibly hard, care deeply, and push me to become a better leader. I’m proud of each and every one of you and the team we’ve built together. We’re solving one of the most important, consequential, and difficult problems in the world. Our vision is massive, and we’re just getting started. Read more here: https://lnkd.in/gxvpHGmW
-
James Detweiler shared thisJames Detweiler shared thisSingularity is solving one of the most urgent industrial challenges of our time: building enough affordable air defense systems to keep pace with modern conflict. Congratulations to Jack Oswald, shail giroux, and the entire team on today’s launch! We at Felicis are true believers in Singularity Defense and are proud to have co-led the Series A.
-
James Detweiler shared thisSingularity is tackling one of defense's biggest bottlenecks: manufacturing affordable air defense systems at scale. In Ukraine, Jack Oswald saw the current state of warfare: inexpensive drones, expensive targets, and too few interceptors. The technology to stop low-cost enemy drones already exists, but the manufacturing capacity to build enough interceptors does not. Enter Singularity. The company is transforming air defense from a scarce capability into one that can be manufactured at industrial scale. Felicis is co-leading Singularity's Series A alongside Khosla Ventures. Jack has built a team to match the mission: his cofounder shail giroux, who sold a robotics company at seventeen before Anduril Industries; VP Todd Morrow, who ran short-range air defense in the US Army and sold billions of it at Lockheed Martin; and engineers from SpaceX, Tesla, Raytheon, Lockheed, Anduril Industries, and ABL Space Systems. Congratulations to Jack, Shail, Todd, and the whole team! More from me on why we backed them: https://lnkd.in/g6xWtjrK
-
James Detweiler reacted on thisWe haven’t had a new energy source that is clean and can run 24/7 in a very long time. Watch this space!James Detweiler reacted on thisWith no established playbook, Endurance approaches subsea geothermal from first principles, building and deploying increasingly complex systems to learn directly from the ocean. Four deployments later, those lessons are coming together in Adélie, the next step toward making subsea geothermal power possible at scale. This is the first film in a series telling the story of what it takes to build something that has never existed before.
-
James Detweiler reacted on thisJames Detweiler reacted on thisOur co-founders, Ekin Dogus Cubuk and Liam Fedus, joined Mario Gabriele on The Generalist to talk about building a synthesis superintelligence, scaling autonomous labs, and why AI's next frontier is engineering physical matter. Full episode here: https://lnkd.in/eh-QSVhW
-
James Detweiler reacted on thisDeployment is the hidden pillar of robotics. I'm excited to share that you can now find Skild robots at the stunning, new Pittsburgh International Airport. This one feels especially meaningful: Pittsburgh is our hometown/HQ, and Skild was founded just down the road at CMU. There’s something pretty special about bringing our robots back to where it all started and putting them to work in one of the most innovative airports.James Detweiler reacted on thisIf you're traveling through Pittsburgh International Airport over the next six months, keep an eye open for its not-so-furry friends – four-legged robot 'dogs' that will roam the airport helping with tasks such as monitoring indoor air quality and detecting spills and rubbish. “Pittsburgh is a global leader in robotics and AI," said Deepak Nayyar, executive VP and CTO at PIT. "We’re looking at how the latest technology can work alongside our staff and airline partners to enhance efficiency, improve operations and provide an additional level of customer service.” Sara Ahmadi, head of product at Skild AI, whose adaptive autonomy software will power the robots, added, “We share the belief that robots will transform how public infrastructure operates.” 🔗 Click the link in the comments to read more! Robert Belle-Isle
-
James Detweiler liked thisJames Detweiler liked thisSupabase began as an open-source alternative to Firebase. Its ambition now is much larger: to become the world’s most successful database company and store all the world’s data. That vision requires solving one of Postgres’s biggest challenges: operating reliably at enormous scale. Through Multigres, Supabase is building an open-source scaling layer designed to bring high availability, operational simplicity, and eventually Vitess-grade horizontal scaling to Postgres. It is a characteristically ambitious next chapter from Paul Copplestone, Ant Wilson, and the Supabase team. They are pairing one of the world’s largest developer communities with deep, long-term technical differentiation. We at Felicis have been true believers since we led Supabase’s Series B in 2022 and are proud to continue supporting the team as they build toward a future where there is no size Postgres cannot reach. This is the story of two founders who built the invisible backend of the AI coding boom without losing the developers who got them there: cc: Viviana Faga Aydin Senkut 🔗 felicis.link/EXia6xQ
-
James Detweiler liked thisJames Detweiler liked thisHuge news from OpenAI DevDay 2026. Baseten has joined the OpenAI Marketplace. Now, enterprise teams can spend their OpenAI commits on Baseten inference, including running open models for code generation natively in Codex. If you're looking for ways to save on tokens, my DMs are open. https://lnkd.in/gutdQ84q
-
James Detweiler liked thisJames Detweiler liked thisAn urgent mission demands an exceptional founder. James Detweiler of Felicis on why Jack Oswald is the one to meet it. Read the full quote below.
-
James Detweiler liked thisJames Detweiler liked thisAI is changing who can build software, and Supabase is the infrastructure beneath that shift. Rather than compete with platforms like Lovable, Bolt, and Vercel, Supabase co-founders Paul Copplestone and Ant Wilson made the astute decision to help power the entire category. As a result, database launches on Supabase grew 600% over the past year, with more than 60% of new databases now started by AI tools. Strategic clarity like that defines enduring companies. Supabase is now indispensable to the AI application ecosystem. We at Felicis have been true believers in Supabase since we led the Series B in 2022. Read the story of how Paul and Ant built Supabase into a $10 billion company while staying true to the developers who got them there: 🔗 felicis.link/rvsh5kW cc: Viviana Faga
-
James Detweiler liked thisJames Detweiler liked thisOne of my favorite things about Supabase is how directly the team speaks to developers. They do not “market” to their community. They participate in it. That includes publishing more than 40,000 memes that reflect the real frustrations, mistakes, and inside jokes of building software. But the memes are more than entertainment. They are a remarkably fast product-feedback loop. When one resonates, the team immediately learns that the problem it names is real and widely felt. That instinct to meet developers where they are has helped Supabase build a community of more than 10 million developers, while staying close to the people who made its growth possible. This Felicis profile on Supabase cofounders Paul Copplestone and Ant Wilson is a great look at the culture and customer understanding behind the company's rise. 🔗 felicis.link/wEut3am
-
James Detweiler reacted on thisJames Detweiler reacted on thisI'm honored to share that Bain Capital Ventures (BCV) is leading Solcoa Industries' $75M seed financing to build Solcoa One, their 500-tonne rare earth metallization plant in Nevada. The origins of American metal production trace back to Alcoa, one of the first domestically innovative production solutions. It exists because Charles Martin Hall invented an electrolytic process that dramatically lowered the cost of producing aluminum in 1886. Solcoa draws on that history: rather than trying to recreate the incumbent rare-earth processes that were industrialized overseas, it is developing a new metallurgical process intended to make rare-earth metals faster, cleaner and cheaper. In just a year, Solcoa has gone from proving its chemistry on a small demo line to producing NdPr and samarium at a scaled facility in Alameda, and now announcing Solcoa One in Nevada. I remember when a furnace vendor quoted us 18 months to deliver, they built the furnace themselves. Demand for rare earths is exploding, driven by EVs, robotics, and every downstream application that runs on permanent magnets, and they meet that need with the same urgency internally. I'm grateful for Hooman and Artem's leadership at this critical time. They bring the ability to fundamentally rethink a process while moving faster than many thought possible.
Experience & Education
Test Scores
-
GRE
Score: 330
Languages
-
English
Native or bilingual proficiency
-
Italian
Limited working proficiency
Other similar profiles
Explore more posts
-
William Webster, CPA
AI and associated computing costs are sending VC-backed CAPEX through the roof Venture debt is picking up the slack - as evidenced by PitchBook reporting Perhaps most interesting is the dislocation in deal counts and deal value post SVB and first republic crashes Anecdotally - we’ve seen a reinvigoration of interest from our tech clients in exploring this financing mechanism Is venture debt having a renaissance? #startups #venturecapital #privateequity
4 Comments -
Keval Desai
There are few VCs who have the barbell experience of having analyzed both public tech companies & invested at the inception stage. It's like having a college professor who's also a kindergarten teacher. Our partner Elizabeth Harrow is such a unicorn. You can see why that's so useful in understanding what's going on today in her conversation with Clare O'Connor at Investor's Business Daily ... cc SHAKTI
1 Comment -
Daniel Dart
🚨NEW EPISODE: Recorded live at FUTURE TITANS 2026 - Jeff Perry of Carta sat down with the iconic Seth Levine, co-founder of Foundry. Seth has been in venture for 25 years, built Foundry from scratch as an emerging manager himself, and has backed about 50 emerging manager funds through his fund of funds. He has genuinely seen every side of this table. They went deep on building Foundry, why VCs are in the influence business, not the decision business, and why the concentration problem in venture is not only bad for LPs, but also for the innovation ecosystem overall. And why Seth's new book, Capital Evolution, is so important for the future of America. 🎧 Links to listen... Apple: https://lnkd.in/ehQUQ2EM Spotify: https://lnkd.in/eU4FExpg
1 Comment
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named James Detweiler in United States
32 others named James Detweiler in United States are on LinkedIn
See others named James Detweiler