Factory Lands $200M to Scale AI-Powered Code Generation Platform

Factory, which deploys AI agents to accelerate software development workflows, announced a $200 million funding round today. The investment was led by Blackstone, Khosla Ventures, Sequoia Capital, NEA and other major institutional investors, alongside angel participation from Marc Benioff. The valuation now stands at $5 billion, up $3.5 billion since April.

The San Francisco AI Factory Inc. operates a platform designed to handle intricate coding assignments, such as converting applications across programming languages. Users describe their desired outcomes in plain language, and the system decomposes these into discrete steps executed by specialized AI agents. The infrastructure required for each agent is provisioned automatically, with built-in safeguards to prevent agents from conflicting with one another. A real-time dashboard provides visibility into project status.

How the Platform Works

Projects begin with a slash command called Readiness Report, which validates whether the code repository is suitable for agent processing. The system identifies repositories that lack proper organization or documentation, which can hinder AI performance. It also confirms that guardrails exist to block agents from introducing security flaws.

Once validation passes, Factory assigns each coding step to the large language model offering the best cost efficiency. Should an LLM encounter problems, the platform automatically redirects the task to an alternative model. After code generation, the platform runs quality assurance and security reviews. A tool named AutoWiki creates accompanying documentation that updates automatically with each code change. Developers can also configure an AI agent to continuously scan the code for potential issues and suggest fixes.

Factory executes agents within isolated virtual environments called Droid Computers, which can run in cloud infrastructure or on-premises systems. Each Droid Computer maintains the libraries, configurations and other technical resources needed by its agents, eliminating the need for manual setup before each use.

Self-Improving Agents

In January, Factory revealed that it uses LLMs to continuously refine its AI agents. These models incorporate improvements derived from anonymized data gathered across thousands of daily sessions.

Across the world's largest enterprises, we are seeing a move from individual coding agents to software factories that serve as the core foundation from which an entire software company operates

Matan Grinberg, Factory co-founder and CEO

Factory disclosed that hundreds of thousands of developers currently use the platform, including teams at major technology companies such as Nvidia Corp., Adobe Inc. and Palo Alto Networks Inc. The company plans to deploy the new capital toward hiring, with a target of 300 employees by year-end.