Salesforce led the Series A, which 8090 announced this week. The pitch behind it is blunt. AI can already write code. The hard part is stopping enterprise software from falling apart as dozens of agents and engineers change it every week. The round was first reported by a tech news outlet, and the company confirmed the details.
The investor list is heavy with names. Alongside Salesforce sit WNDR, Craft Ventures, The Production Board, and LAUNCH. Angels include Nikesh Arora, Cliff Robbins, Adam D’Angelo, and Thomas Laffont. The capital will go on hiring and on the compute needed to run the product at scale.
What 8090 actually sells
8090 calls its product a “software factory”. The idea is a single governed workspace where people and AI agents build and change enterprise software together. It tries to connect the whole chain, from business intent and requirements through architecture, code, testing, and production upkeep.
The selling point is not raw speed. It is control. 8090 promises leaders visibility, accountability, and an audit trail from idea to deployment. It aims that promise straight at what makes big companies nervous about AI. The worry is not whether AI can write code. It is whether anyone can see what it changed, and why.
The company also runs a delivery arm. It designs, builds, hosts, and maintains custom systems for clients in regulated industries. Those include healthcare, insurance, life sciences, manufacturing, financial services, and government. That work, 8090 says, hardens the platform against messy legacy systems.
The numbers it is leaning on
8090 backs the pitch with a set of customer results. They are worth repeating, with one caveat: these are the company’s own figures, not independently checked.
By its account, 8090 turned more than 18 million lines of COBOL and Assembly into plain English. That code sat behind a healthcare billing engine. It became over 300,000 readable rules in 40 days. The company says a listed health insurer then cut claims sent to a pay-per-catch vendor by 80 per cent, avoiding more than $20M over four years. A life sciences customer cut a diagnostic’s time to market from five years to four, it adds. A manufacturer brought more than 10,000 parts under real-time validation.
If those results hold up outside the press release, they point at the real prize. It is not greenfield apps. It is the expensive, brittle systems that large firms cannot easily replace.
Why Chamath in the chair matters
The headline is not really the money. It is the job. Chamath Palihapitiya has spent the years since Facebook as an investor and a prolific SPAC sponsor. That period left him rich and loud. For many retail investors who bought into those deals, it was a mixed bargain, and several of his SPAC targets fell hard after listing.
So a return to a full-time operating role is a statement. “Since I left Facebook, I was waiting for a moment like this to return to a full-time operating role,” he wrote in a blog post announcing the raise. He framed AI as “the grand equaliser” and said the next few years would set the stage for the next twenty.
That is the optimistic read. The sceptical one is simpler. Founder-investors with strong brands raise large rounds on narrative as much as traction, and 8090 is young. Salesforce putting its name at the top of the round buys credibility the founder’s reputation alone would not.
Chamath’s background adds layers. He was an early Facebook executive, joining in 2007 and leading the growth team until 2011. After leaving, he founded Social Capital, a venture firm that invested in companies like Slack and Ring. He later became famous for SPACs—special purpose acquisition companies—that took firms like Virgin Galactic and Clover Health public. While some deals soared initially, many struggled post-merger, drawing regulatory scrutiny and shareholder lawsuits. Critics say Chamath’s SPAC era prioritized hype over sustainable value. His return to an operating role signals a pivot from being merely a financier to a builder. He is betting his reputation on 8090’s ability to deliver real enterprise outcomes.
How 8090 fits into the AI coding frenzy
8090 is landing in the middle of a funding frenzy. Investors keep pouring money into AI coding and agent startups, even as the cost of running these tools climbs. The demand is real, and AI labs are winning paying customers fast. Enterprises already feel that squeeze. Amazon is among those hunting for cheaper alternatives.
The same pressure is pushing rivals to build in-house. Meta has gone as far as restricting its engineers’ use of Anthropic’s Claude Code and OpenAI’s Codex while it builds its own tool. 8090 wants to sit one layer up. It sells the orchestration and oversight, not the model. That is the same governance-first instinct now drawing money into agentic security.
Startups like Cognition’s Devin, Replit, and GitHub Copilot are targeting developer productivity. But 8090 differentiates by focusing on enterprises with legacy systems, compliance needs, and multi‑agent chaos. It argues that the real bottleneck isn’t generating code—it’s keeping production stable as changes accumulate. This governance layer is reminiscent of what DevOps and CI/CD pipelines did for traditional software, but adapted for an AI‑driven world.
The legacy system opportunity
One of 8090’s biggest selling points is its ability to modernize old code. Many large corporations still run core systems on COBOL, Fortran, or Assembly. These languages are decades old, and the engineers who wrote them are retiring. Banks, insurers, and government agencies face a ticking clock: either modernize or risk system failure. 8090’s approach—translating legacy code into readable business rules—offers a path that doesn’t require full rewrites. The company claims it can do this in weeks, not years.
The healthcare billing example shows the potential. Transforming 18 million lines of COBOL into 300,000 rules gave the client an unprecedented view of their claims process. The 80% reduction in vendor‑sent claims suggests that hidden waste gets exposed when code becomes human‑readable. Similarly, cutting a diagnostic’s go‑to‑market from five years to four in life sciences accelerates revenue and patient access. For manufacturing, real‑time part validation reduces defects and recalls.
These outcomes, if consistent, could unlock a massive market. Consulting firm Gartner estimates that global spending on legacy application modernization will exceed $300 billion by 2027. 8090 positions itself as a tool that speeds up that process while maintaining governance. Competitors like Micro Focus and IBM offer modernization tools, but 8090 claims its AI‑first, agent‑driven methodology is faster and more transparent.
The governance imperative
Beyond legacy conversion, 8090’s core thesis is about controlling the chaos of AI‑generated changes. In a typical enterprise, dozens of software agents—each trained to write code or fix bugs—operate in parallel. Without a central orchestration layer, conflicts, regressions, and security holes multiply. 8090 provides a workspace where every change is tracked, reviewed, and deployable only with full audit logs.
This aligns with the broader trend of “agentic security”—tools that monitor and control AI agent behavior. Companies like WriteSonic, Lasso Security, and Patronus AI have raised venture capital for similar reasons. Enterprises want to adopt AI without losing control. 8090 essentially offers a pair of guardrails for code generation, a layer that sits between the model and the production environment.
The governance feature also appeals to regulated industries. Healthcare, financial services, and government must comply with rules like HIPAA, SOX, and GDPR. 8090’s audit trail provides evidence that changes were authorized and tested. That could reduce the risk of regulatory fines and lawsuits. The company’s own delivery arm, which builds and maintains systems for such clients, serves as a testing ground for the platform’s features.
What the competition looks like
8090 isn’t the only AI startup chasing the “software factory” concept. Others include GitLab’s AI features, Codeium, and Sourcegraph’s Cody. But most focus on improving developer speed. 8090 instead targets the manager, the CTO, the compliance officer. It sells a control panel, not a code editor.
Meanwhile, cloud giants are moving in. Amazon Bedrock allows enterprises to build custom coding agents. Google Cloud’s Vertex AI offers similar tools. Microsoft’s GitHub Copilot already has enterprise plans with governance controls. 8090’s advantage lies in its deliberate focus on orchestration and legacy systems, areas that the hyperscalers often treat as add‑ons rather than core products.
The funding round—a $135M Series A—is unusually large for that stage. Typical Series A rounds average $10‑20 million for enterprise SaaS. The size reflects both the capital needed for compute and the investor belief that 8090 can become a category champion. Salesforce’s involvement is particularly strategic. The CRM giant is pushing its own AI platform, Einstein, and sees value in a partner that can manage the code delivery pipeline for its largest customers.
The open question is whether “a factory for agents” is a product or a slogan. Plenty of firms can wire a coding model into a workflow. 8090 is betting that the hard, unglamorous parts are where the durable business sits: the governance, the audit trail, the legacy systems. With Chamath now staking his own time on it, the company will not lack for attention. It will have to earn the rest.