Jason
Warner

Most people who incubate a world-changing product inside a big company collect the bonus and write the LinkedIn post. Jason Warner left GitHub, where he ran engineering as CTO, having watched a small experimental project called Copilot go from internal skunkworks to something that rewired how every developer on earth thinks about writing code - and decided the interesting work was still ahead of him, not behind.

In April 2023, Warner co-founded Poolside with Eiso Kant - the two had first met in 2017 and spent six years quietly mapping the territory before planting their flag. Their thesis is narrow and deliberate: software development will be the first broad domain where AI surpasses human-level intelligence. Not creative writing, not medical diagnosis - code. And the path there runs through reinforcement learning from real code execution feedback, not just predicting the next token.

The CV that gets him taken seriously: before Poolside, Warner was the Managing Director at Redpoint Ventures on the early growth team, backing developer tools and infrastructure companies. Before Redpoint, he was CTO at GitHub for four years - where, alongside managing the platform strategy, he brought GitHub Actions, Packages, Advanced Security, Connect, and Codespaces to market. And before GitHub, he was VP of Engineering at Heroku and led product engineering for Ubuntu Desktop and Ubuntu Phone at Canonical. The thread across all of it: infrastructure, developer experience, and the machinery of how software gets made at scale.

Warner studied computer science at Penn State for his bachelor's and then at Rensselaer Polytechnic Institute for a master's. He describes himself, without irony, as "a very average programmer but an excellent, excellent architect - an excellent engineer." It's a distinction that matters. Average programmers write functions. Architects design systems. And Warner's career has been a long argument that the system - the organization, the toolchain, the feedback loop - is where the leverage lives.

At GitHub, his formative experience was both technical and organizational. He saw Copilot, the AI coding assistant that would become Microsoft's billion-dollar showcase, incubate quietly in his organization - and recognized that the underlying shift it represented was only beginning. GitHub Copilot autocompletes. What Warner is building at Poolside reasons about code, runs it, learns from the execution results, and iterates. The difference is roughly the gap between a GPS giving you directions and a co-pilot who has driven that route a thousand times and knows which turns look right but aren't.

Warner has a clear and somewhat contrarian view on who should be building foundation models. His answer: almost nobody. Most companies, in his telling, should be building applications on top of existing models - that's the sensible, defensible business. Poolside builds foundation models precisely because Warner believes intelligence will become the world's most critical resource, comparable to electricity, and only the companies who treat it that way - not as a fundraising vehicle but as a transformative commodity - will be worth building from scratch.

The Series B, closed in October 2024, brought in Bain Capital Ventures as lead, alongside NVIDIA, Citi Ventures, Capital One Ventures, HSBC Ventures, LG Technology Ventures, and eBay Ventures. The round valued the company at $3 billion. By October 2025, Nvidia announced it would invest up to $1 billion in Poolside - a move projected to push the valuation to roughly $12 billion. The company had 10,000-plus Nvidia GPUs running full-tilt on model training.

The ambition has moved into physical infrastructure too. Project Horizon - a two-gigawatt AI data center campus in Pecos County, Texas - launched its first phase in October 2025. Poolside simultaneously signed contracts with RTX Corporation and companies in the American defense industrial base, established a strategic partnership with AWS for AI deployment on Amazon Bedrock, and maintained its hybrid presence between San Francisco and a Paris office where concentrated monthly work weeks happen. For a company Warner describes as being "forged at this very moment," the pace is consistent with the rhetoric.

There's a recurring mental model in how Warner talks about teams and companies. He views human organizations as "lossier versions of computer distributed systems" - the same fundamental challenges of coordination, latency, consistency, and failure modes, just running on biological hardware with worse error correction. It's a frame that makes him unusually precise about culture.

Early in his management career, someone told Warner his personality was better suited for marketing or sales. He initially rejected the idea - then realized the insight was actually pointing somewhere useful: engineering leaders who think like CEOs, who can translate between technical depth and strategic vision, are genuinely rare and genuinely valuable. He built toward that rather than away from it.

One story he tells: he once managed a highly talented but deeply toxic engineer. The team worked around that person, absorbed the friction, and treated the situation as fixed. When Warner removed the engineer from the team, productivity went up significantly. The lesson he drew wasn't about bad apples - it was about the invisible tax that weak culture forces everyone to pay, silently, continuously.

He also declines to mimic. "You can't copy," he's said of the instinct to reverse-engineer what Jobs or Musk did. "You have to learn. You need to look at the context of the situation." It's the difference between an architect who studies Gehry and one who pastes Gehry onto a different building. One is education; the other is cargo cult.

Warner's definition of AGI is deliberately functional: "when an AI can basically do the equivalent information work for a person." No mysticism, no Terminator scenarios. An AI that handles the tasks a knowledge worker handles. He believes Poolside's approach - reinforcement learning from code execution feedback - puts software in reach of that threshold before any other domain.

His case for why this matters for workers is equally measured. Companies will face a choice: use AGI to reduce headcount, or use it to magnify what each person can do. Warner argues for the latter - not on ethical grounds alone, but on competitive grounds. A company that keeps expanding human capability with AI can grow faster and build better than one that simply trims. His concrete image for this is an infinitely patient senior colleague available to every new hire: someone who has read the entire codebase, remembers every architectural decision, and will explain any tradeoff without sighing.

Poolside joined the Atlassian Board of Directors in September 2025 - effective October 1 - with Warner appointed directly. He also sits on the Operating Board of Bridgewater Associates, the world's largest hedge fund. The breadth is intentional: financial services, enterprise software, defense, developer tools. Poolside's customers and partners span all of them. Warner is building a company that looks less like a startup and more like foundational infrastructure for a post-AGI software industry.

He once said the only thing he wakes up thinking about is what's next. At Poolside, "what's next" means 2-gigawatt data centers in Texas, billion-dollar Nvidia investments, and a theory that the first machine to truly outthink a human will do it in a text editor. The average programmer turned GitHub CTO turned VC turned AI founder is still running forward, still looking at what's coming. The pace hasn't changed. Only the stakes have.