Factory Raises $200M at $5B for Enterprise AI Agents

Writing code is the visible trick. The enterprise contract lives in the system that decides which model works, where an agent can run, what it is allowed to touch, and whether the output improved anything.

Factory has raised $200M at a $5B valuation to expand that system. The September 15, 2026 financing arrived 5 months after the company's $150M Series C at a $1.5B valuation, pushing Factory's reported total funding above $400M while enterprise AI moves from individual assistants toward governed software production.

What Factory Raised and Who Participated

The new financing includes Blackstone, Khosla Ventures, Sequoia Capital, Insight Partners, Evantic Capital, Sound Ventures, NEA, Mantis VC, and Clearlake. Factory did not identify a lead investor or label the round, so calling it a Series D would add a fact the announcement does not provide.

Factory said it will use the capital to accelerate research, product development, and global go-to-market. The company also disclosed a $5B valuation, more than 3 times the $1.5B valuation attached to its April 2026 Series C, but did not specify whether the new figure is pre-money or post-money or disclose ownership, terms, or board changes.

The Five-Month Valuation Reset

Five months can disappear inside an enterprise software evaluation. Security review, procurement, data architecture, legal approval, deployment design, and internal workflow changes rarely move at startup speed, especially when an AI system can write and execute code across production environments.

Factory's capital timeline moved differently. Its April round brought in $150M and valued the company at $1.5B; the September financing added $200M and valued it at $5B. That acceleration says more about investor expectations for the enterprise control layer than it does about coding demos, because the harder commercial problem begins after an agent produces a convincing patch.

From Droids to a Software Factory

Matan Grinberg and Eno Reyes founded Factory in 2023 around autonomous software-engineering agents called Droids. Matan Grinberg serves as co-founder and CEO, while Eno Reyes is co-founder and CTO. Their original product story focused on agents that could plan, write, test, review, and document software work.

Factory's current argument is broader. In Factory 2.0, the company describes a continuous system that begins with business requirements, bugs, conversations, and customer feedback, then moves through planning, building, testing, security, deployment, and monitoring. The feedback from production becomes input for the next cycle, turning the platform from a collection of tasks into operating infrastructure for engineering.

Control Becomes Part of the Product

Enterprise buyers do not evaluate autonomy only by how much code an agent can generate. They also care about where the system runs, which models it uses, how permissions are enforced, what evidence is available for review, and whether the company can keep sensitive work inside its own security boundary.

Factory supports cloud, on-premises, and fully air-gapped deployments. It also offers task-level model routing, giving customers a way to choose among models based on cost, speed, and performance instead of committing the entire organization to one provider. Factory says its Router reduced token spending by more than 60% while maintaining frontier performance; that figure is company-reported, but it explains why model selection is becoming a budget function as well as a technical one.

Customers Are Buying the System Around the Agent

Factory says hundreds of thousands of developers use its platform and names Nvidia, Blackstone, Royal Bank of Canada, Palo Alto Networks, Adobe, and T-Mobile among the enterprises building software factories with it. Those relationships span technology, finance, cybersecurity, and telecommunications, markets where deployment controls and measurable outcomes matter as much as raw agent capability.

The company has also introduced Agent Effectiveness to help enterprises connect AI spending with measurable value. That measurement layer matters because engineering leaders are being asked to justify both the cost of models and the organizational disruption required to let agents operate across more of the software lifecycle.

The Competitive Line Is Moving

The coding-agent market is crowded because code is one of the clearest places to demonstrate what current models can do. Enterprise differentiation is moving toward everything around that demonstration: access controls, model choice, repeatable workflows, deployment options, production evidence, and the ability to improve the system without surrendering organizational context.

Factory is betting that customers will treat those pieces as one operating layer rather than assemble them from separate tools. That choice raises the commercial stakes. A broader platform can support larger contracts and deeper adoption, but it also has to survive more security reviews, integrate with more systems, and prove value across more of the engineering organization.

What the $200M Changes

The financing gives Factory room to deepen the product while expanding the forward-deployed and go-to-market work required to make agentic systems function inside large organizations. Research can improve the agents, but adoption still depends on integration, governance, workflow design, and the slow accumulation of trust across engineering, security, legal, and executive teams.

Investors are valuing Factory as though that control layer can become a durable enterprise category. The company now has to carry a $5B expectation into organizations where software changes create operating consequences, and where each additional autonomous step has to arrive with clearer evidence, tighter permissions, and a result somebody can defend.