Chamath Palihapitiya is putting his own name back on an operating job, and the bet is bigger than another AI coding tool.
8090 Labs, the AI coding company Palihapitiya founded in January 2024, says it has raised a $135 million Series A led by Salesforce Ventures. The check is not the whole story. Palihapitiya is also stepping in as full-time CEO, which makes this his first day-to-day operating role since he left Facebook in 2011.
That matters because Palihapitiya has spent much of the past decade as an investor, SPAC sponsor and media figure, not as the person responsible for shipping enterprise software to customers who call when it breaks. You can talk about AI coding agents on a podcast. Running one inside a bank, a healthcare company or a consultancy is a different job.
The round also included WndrCo, the firm run by former Disney and DreamWorks executive Jeffrey Katzenberg, along with David Sacks' Craft Ventures, David Friedberg's The Production Board and Jason Calacanis's Launch. Angels included Palo Alto Networks CEO Nikesh Arora, Quora CEO Adam D'Angelo and Cliff Robbins.
Three of those venture backers are Palihapitiya's co-hosts on the All-In podcast.
That's not a normal cap table. It's closer to a group chat with money attached. Frankly, it also means the next test can't be whether friends around the table like the pitch. It has to be whether customers keep using the product after the first deployment stops feeling new.
What 8090 Actually Sells #
8090's product, Software Factory, is not pitched as another autocomplete box inside a developer's editor. The company says it manages a project across requirements, architecture, planning and code execution, with a persistent knowledge graph that builds context as the work moves. That is the part worth watching. It is selling to the organisation. Not the engineer.
That puts 8090 in a different lane from Cursor and GitHub Copilot, which became popular by sitting close to individual developers and answering prompts quickly. 8090 is making a heavier claim: that large companies need an AI development system with memory, auditability and enough structure for production software, not just faster snippets.
The company says it has signed EY, Italian pharmaceutical company Dompé, home healthcare provider AdaptHealth and clean energy platform Palmetto as customers. The EY relationship is the most important one named so far. In March 2026, EY launched EY.ai PDLC, an internal delivery platform built on 8090's Software Factory, with plans to roll it out to tens of thousands of consultants.
That is a real proving ground. Consultants don't need another demo video. They need a system that can survive messy requirements, legacy code, client reviews and the usual last-minute change that arrives after everyone already agreed the work was finished.
Palihapitiya has been direct about the market he wants. In a post on X, he said 8090 works on production systems for large, often regulated enterprises and argued that casual vibe coding is not tolerated in banking, power, healthcare and insurance. The tone is blunt. The claim is also measurable. Either Software Factory can handle that kind of work, or it can't.
The Hard Part Starts After The Raise #
The AI coding market around 8090 is already moving at a ridiculous pace. The Associated Press reported in June 2026 that SpaceX agreed to buy Cursor maker Anysphere in an all-stock deal valued at $60 billion. Business Insider reported in March 2026 that Palihapitiya had already complained on the All-In podcast that 8090's AI costs had more than tripled since November 2025, with annual AI spending heading toward $10 million.
So the pressure on 8090 is not abstract. Model costs matter. Procurement matters. So do integrations and security reviews - and the long list of boring things that decide whether enterprise software lives or dies after the sales deck has done its work.
Palihapitiya's record makes that scrutiny fair. He built Social Capital after leaving Facebook, then became one of the public faces of the SPAC boom through deals involving Virgin Galactic, Clover Health and Opendoor. Some of those companies became famous for the wrong reasons after the market cooled, and Palihapitiya later acknowledged the incentive problem in SPACs, saying he could do a deal and still get paid.
This time he doesn't get to stand one step away from the operating mess. Software Factory is now his to defend. Enterprise customers judge software in procurement meetings, but they really judge it later, during outages, migrations, compliance checks and uncomfortable calls with the people who signed the contract.
The investors backing this round are wagering that Palihapitiya can move from commentator back to builder. EY's rollout gives him an early stage big enough to matter. Regulated industries don't forgive broken software because the seller is famous, and they don't keep paying for systems that make work harder than the old way.
That is the bet inside the $135 million raise. Not whether AI can write code. Everyone already knows it can. The harder question is whether 8090 can make AI software development boring enough, documented enough, auditable enough and reliable enough for companies that can't afford to vibe their way through production.
Also read: StepFun Unveils the First Agentic AI Phone Ahead of Apple and OpenAI • Reflection AI Secures Over $1 Billion in Nvidia Compute From Nebius • GMI Cloud seeks $635 million loan backed by its Nvidia GPU contracts