Factory hits $1.5B valuation to take on Cursor and Claude

The AI coding gold rush is far from over. Even with heavy hitters like Anthropic's Claude Code, Cursor, and Cognition fighting for developers' attention, top venture capitalists clearly believe there is room for another multi-billion-dollar contender.

Factory, an AI startup building autonomous software engineering agents for enterprise teams, just announced a $150 million funding round at a $1.5 billion valuation. The round was led by Khosla Ventures, with participation from Sequoia Capital, Insight Partners, and Blackstone. As part of the deal, Khosla managing director Keith Rabois joined Factory's board.

From physics PhD to billion-dollar founder

Factory was founded in 2023 out of a cold email. Matan Grinberg, then a PhD student at UC Berkeley, reached out to Sequoia partner Shaun Maguire.

It turned out Maguire held a Caltech PhD in the exact same physics subfield Grinberg was studying. The mutual academic connection led Maguire to convince Grinberg to drop out and start Factory, with Sequoia cutting the check for their seed round.

More than three years after generative AI entered the mainstream, automated coding remains the single most lucrative and validated use case across the industry. That tailwind helped Factory scale its valuation at breakneck speed.

Courting Morgan Stanley with multi-model freedom

While rivals battle for individual programmers, Factory is laser-focused on enterprise infrastructure. Its active customers already include engineering departments at Morgan Stanley, Ernst & Young, and Palo Alto Networks.

Grinberg told The Wall Street Journal that Factory's key differentiator is letting developers seamlessly swap foundation models under the hood, routing tasks to Anthropic's Claude or Chinese startup DeepSeek depending on the job. That said, competitors like Cursor also avoid locking users into a single model family.

Why it matters

We think Factory's rise signals a major shift in enterprise software procurement. Large corporations are terrified of vendor lock-in, and tying their entire code pipeline to one proprietary frontier model feels dangerous.

By positioning itself as model-agnostic, letting enterprises tap Claude or DeepSeek interchangeably, Factory hits a sweet spot for security-minded CIOs. Winning Wall Street stalwarts like Morgan Stanley and accounting giants like EY shows that enterprise IT buyers want an orchestrator they control, not just a chatbot inside an editor.

Still, technical differentiation remains Factory's biggest question mark. Switching models is a feature, not a moat; rivals like Cursor already offer multi-model support. Factory's real test won't be whether it can call different APIs, but whether its enterprise-grade permissions, compliance, and agent workflows can keep corporate clients from switching when the next big IDE comes along.