Jeff
Wang

The Call No One Expected

On a Tuesday in July 2025, Windsurf's founders - Varun Mohan and Douglas Chen - pulled Jeff Wang aside. Google was hiring them away. A licensing deal, valued at $2.4 billion, was weeks from closing. The previously announced $3 billion OpenAI acquisition had already quietly collapsed. By Friday, Windsurf's original leadership would be gone.

Wang - who had spent the previous two years building Windsurf's entire go-to-market engine as Head of Business - had 48 hours to process the news before telling 250 employees that their expected payday had evaporated and the people who built the company were leaving.

What Wang did next is the reason Windsurf still exists as a coherent entity. Rather than managing a slow unwind - competitors circling, engineers fleeing, customers spooked - he made a single decisive call: find an acquirer. Not tomorrow. That Friday.

By Monday morning, he had one. Cognition, the company behind Devin (the AI software engineer), agreed to acquire Windsurf's remaining team and technology. One lawyer involved described it as among the fastest deals they'd ever handled. Jeff Wang had lost 8 pounds by then and gotten roughly three hours of sleep across the weekend.

Before the Crisis, the Grind

Wang grew up in Chicago. His father, an engineer himself, gave him one piece of career advice with an unusual caveat: you can be anything you want, as long as it's an engineer. Wang obliged - earning a B.S. in Electrical Engineering from the University of Michigan in 2008, then spending years at large tech firms that started feeling too big. Cisco. Then Salesforce, where he was a Program Manager while simultaneously completing his MBA at UC Berkeley's Haas School of Business.

That degree finished in 2020, he bounced through a sequence of early-stage bets: Solutions Architect at Tonkean, VP of Product (Crypto) at TomoCredit, Entrepreneur in Residence at Workplay Labs. He was also quietly building RNR Capital as a General Partner, placing angel checks in startups several of which would grow to unicorn scale. He'd co-founded RocketFuel Education, teaching people the mechanics of crypto, AI, economics, and venture. He was accumulating surface area, not climbing a single ladder.

He found Codeium - later renamed Windsurf - through a Kleiner Perkins partner's LinkedIn post seeking a first business hire for an AI coding startup. Wang had been personally tinkering with AI since GPT-2 dropped. He'd been running Llama models at home, experimenting with diffusion models. When he met founders Varun and Douglas, the pattern matched immediately: smart team, genuine technology, right timing.

He joined in 2023 as the company's first business hire. No sales team. No marketing playbook. No CRM beyond Google Docs and Sheets.

500 Customer Calls. Thirty-Minute Intervals.

What followed was a relentless, deliberately low-tech discovery process. Wang and colleague Ancho ran 12-hour days booking one new customer conversation every 30 minutes, continuously A/B testing their messaging in real time. They were hunting for something specific: not prospects who were mildly interested in AI coding tools, but people whose problem was genuinely painful.

They found it by zigging where Copilot zagged. GitHub Copilot, backed by Microsoft, had locked in cloud-hosted development for mainstream enterprises. Wang's team targeted the overlooked edge: companies using Bitbucket or GitLab instead of GitHub, regulated industries that couldn't send code to Microsoft's servers, and organizations requiring on-premises deployment. It wasn't a glamorous niche. It was an undefended one.

First deal: $1,000. Three to four months later: seven-figure contracts. Wang built the GTM playbook by living inside every customer conversation, then systematized what he learned into internal playbooks - using AI, eventually, to do account research, stakeholder mapping, and business performance analysis at scale. He was practicing what Windsurf was selling.

Within the company, he also intentionally made himself replaceable at each stage - moving from marketing to finance to sales to solutions engineering to operations - building the infrastructure behind him before stepping aside into whatever gap had opened next.

The Deal, and What It Says About the People

The Cognition acquisition happened because it made structural sense. Windsurf had built the leading in-IDE agentic coding experience - the kind of tool a developer uses in real-time, inside their workflow. Cognition's Devin operates asynchronously, handling longer-horizon engineering tasks in the background. The products don't compete. The teams didn't overlap. Cognition had overinvested in engineering and had almost no go-to-market infrastructure. Windsurf's GTM function, which Wang had built from scratch, was - in his words - "nothing short of world class."

What made the deal unusual wasn't the fit. It was the speed and the terms. Wang negotiated full vesting acceleration for all Windsurf employees, waived cliffs, and equity payouts across the board. When he told the team what had been secured - "We've decided to give you one year of vest, OH, and years two, three, and four as well" - the applause, he said, seemed to last forever.

Asked what he thinks employees will remember him for, Wang's answer is disarmingly quiet: "I actually don't know. I think they will just think we did the right thing."

What He Thinks Is Coming

Wang sees AI-assisted software development approaching an inflection point that most people are still underestimating. His projection: 10x to 100x productivity gains, with the volume of code written annually exploding - not because the same developers write more, but because non-developers start contributing to software creation. Legacy systems modernization alone represents an enormous untapped backlog. Every global company has software that needs to be built or rebuilt.

He's explicit about the longer arc: the IDE itself will eventually become less central. Agentic tools - autonomous systems that can take a goal and execute on it across time - will increasingly handle what humans now manage manually. The engineer's job shifts toward delegation, management, and architectural judgment.

His private shorthand for the plan is "diabolical": build a $100 billion business by enabling every organization on earth to build software at speeds previously unimaginable. Cognition's Devin and Windsurf's agentic IDE are, in Wang's view, complementary tools for exactly that future.

The Investor, the Educator, the Tinkerer

Beyond Windsurf, Wang remains a General Partner at RNR Capital and an active angel investor - multiple portfolio companies have grown to unicorn scale. He co-founded RocketFuel Education to teach the mechanics of crypto, AI, economics, and venture capital to people who weren't yet inside these systems. He writes at jeffwang.substack.com and maintains a Twitter presence at @jeffwsurf, where he was posting about Windsurf and AI coding well before the acquisition drama made it everyone else's story.

He publicly calls Cognition's Scott Wu "maybe the most underrated CEO right now in Silicon Valley" - citing Wu's ability to assemble a team that includes multiple International Olympiad in Informatics gold medalists and his execution velocity in a market with minimal established competition. Wang's admiration for Wu is specific, unhedged, and doesn't read like networking. It reads like recognition between people who both understand what they're building.

Wang joined Twitter in February 2009 - back when social media felt like a new idea. He's been watching technological shifts accumulate longer than most people in the AI coding space have been paying attention. His Twitter handle - @jeffwsurf - merges his name with Windsurf, the company he helped build and then saved. It was probably the fastest personal rebrand in the history of a crisis.