Factory, an enterprise AI coding platform, raised $200m on 15 September 2026 at a $5bn valuation — more than triple the $1.5bn it was valued at just five months earlier, in April. The round was led by Blackstone, Khosla Ventures, Sequoia Capital and NEA, with Salesforce CEO Marc Benioff investing personally. The company says its platform is already used by hundreds of thousands of developers, including staff at Nvidia, Adobe and Palo Alto Networks.
What Factory sells is a step beyond the single-agent coding assistant most founders have already heard of. Its “Droids” are autonomous agents that take a high-level task description, break it into sub-steps, and hand pieces to specialised agents that write code, provision infrastructure, run quality and security checks, and generate documentation — persisting their environment and configuration between sessions rather than starting cold each time. Factory’s own framing for this is a shift from “individual coding agents” to “software factories that serve as the core foundation from which an entire software company operates.”
Why the valuation jump matters more than the product
A 3.3x valuation increase in five months, on top of Cognition’s Devin nearly doubling to a $47bn valuation over the same stretch, tells you where investor and enterprise money is actually flowing right now: not into better autocomplete, but into agent platforms that own multi-step engineering workflows end to end. That’s a meaningful signal for anyone evaluating what “AI-assisted development” means for their own project in 2026 — the category has moved well past a developer typing prompts into an editor.
So what
The capability gap between a single AI coding assistant and an orchestrated agent platform is real, but so is the gap between what a platform like Factory can autonomously produce and what’s actually ready to ship to production, paying customers, or regulated environments. The tools are getting more powerful faster than most buyers can evaluate them — which is exactly where a development partner earns their keep: knowing which parts of this stack to lean on, and which still need an experienced engineer reviewing every line. See our AI-assisted development approach, or get in touch if you’re trying to work out what these tools actually change for your build.