Chamath Palihapitiya has closed a $135 million Series A for 8090, an AI coding startup he founded in January 2024, and has assumed the CEO role himself, according to TechCrunch.

8090's thesis targets a structural weakness in enterprise software: rebuild applications to 80% feature completeness and deliver them at 90% lower cost than legacy systems. The company name encodes the value proposition directly. Most enterprise buyers never use the long tail of features in mature ERP or CRM systems (ERP and CRM are enterprise resource planning and customer relationship management platforms). 8090 bets that an AI-native build focused on core functionality can hit switching economics strong enough to pull customers away from incumbents, per Forbes.

The company is not cutting corners on infrastructure. 8090 is on pace to spend more than $10 million annually on AI costs alone — a figure reported by Business Insider in March 2026. At that burn rate, the cost of running AI inference (the computational work of processing requests through a trained model) and licensing models become a real line item in the unit economics — not a rounding error. That pressure directly tests whether the 90%-cost-reduction claim holds as the product scales.

The Bet Against Bloat

The 80/90 framing is not new. Enterprise software vendors have faced "good enough" competitors for decades: Salesforce undercut Siebel on implementation complexity; vertical SaaS startups carved niches from SAP and Oracle. What shifts now is the mechanism. AI-assisted code generation compresses the time and headcount required to reach feature parity on an application's functional core. A startup that might once have needed five years and 200 engineers to build a credible accounts payable module can, in theory, accomplish it faster with a smaller team.

Palihapitiya stepping into the CEO seat himself, rather than hiring an operating executive, signals a personal conviction bet. His primary public identity remains venture capitalist and co-founder of Social Capital; the operational move suggests he expects to spend serious time on 8090.

The $135 million Series A is a substantial opening round for enterprise software, arriving at a moment when AI coding tooling has drawn intense capital — from Anysphere's Cursor to GitHub Copilot to a growing roster of AI agent-based entrants. 8090's stated differentiation is not the coding tool itself but the finished enterprise application: selling outcomes to procurement teams, not capabilities to individual developers.

The 80% feature completeness figure deserves a closer look. In regulated industries — financial services, healthcare, government — the "missing" 20% often contains exactly the compliance, audit trail, and integration requirements that make a system viable in practice. Whether 8090's feature math holds up under enterprise security reviews and procurement due diligence is an open question. The Series A will help the company test that assumption, but not guarantee the answer.

The $10 million-plus annual AI cost also warrants attention as a leading indicator of future unit economics. If inference costs continue falling — driven by model efficiency gains and competition among cloud providers — that figure could shrink as a percentage of revenue. If 8090's architecture is tightly coupled to today's model pricing, the company gets structural improvement as markets change. That is a meaningful tailwind, assuming the product ships at scale.

Palihapitiya has spent years arguing that most enterprise software is over-engineered and over-priced. With 8090, he is now running the operational test of that argument himself, with nine-figure capital and his name on the door.