Prime Intellect raises $130M in Series A to help companies build their own AI agents

Prime Intellect, a startup that provides computing power and specialized software tools to help companies build AI agents, has raised $130 million in Series A at a $1 billion valuation.

This large round was led by Radical Ventures, with participation from Nvidia Ventures, Intel Capital, Dell Technologies Capital, Iconiq, and a long list of angel investors who are founders of notable companies, including Aravind Srinivas (Perplexity), Aaron Levie (Box), Winston Weinberg (Harvey), Jeff Wang (Cognition), and Brendan Foody (Mercor).

Founded in 2024, Prime Intellect’s goal is to provide organizations with the ability to train their own agent systems without relying on state-of-the-art AI labs. Achieving this mission would have been difficult just a few years ago, but the rise of reinforcement learning techniques that repeatedly reward task success and penalize errors has enabled companies to become their “own AI labs” by refining models for specific business tasks.

Although it is now possible to avoid closed AI labs, the underlying infrastructure is still so complex that most companies do not have the expertise to assemble these pieces into production-ready systems.

That’s where Prime Intellect comes into play.

The startup has developed what it calls a “full stack” for AI agent development, including compute access, reinforcement learning frameworks, and evaluation tools.

Prime Intellect’s platform works like a marketplace, offering modular access so customers can choose the specific tools they need without being locked into an all-or-nothing system.

“They’ve pieced this together and built it in a way that allows them to operate at the forefront at an affordable price,” said David Katz, partner at Radical Ventures. He added that Prime Intellect is unique in offering top-level AI lab capabilities as a “one-stop shop” for development, while other companies offer piecemeal offerings.

The startup’s approach has attracted customers like Ramp, Zapier, and Flapping Airplanes, who pay the startup for hosted versions of the tool. This rapid adoption helped the company reach $100 million in annual sales.

This growth is driven by tangible results. For example, Ramp used Prime Intellect to build an agent that helps fintechs find answers within spreadsheets. “The results outperformed Frontier models in accuracy while running faster and at a fraction of the cost,” Karim Atiyeh, co-founder and co-CEO of Ramp, said in a statement.

Another key factor driving Prime Intellect’s growth is the company’s recent recognition that building on top of Frontier Labs comes with many risks.

Companies increasingly don’t want to provide their sensitive information to OpenAI or Anthropic due to the risk of losing control of their data. I’m also wary of relying on models that can suddenly power off, like Anthropic’s Fable did last month.

“How do I know I’m not working for a company that’s going to replace me and popularize what I do?” Katz said. “All of this has led people to think, ‘How can I own my own enterprise intelligence without taking on these risks?'”

Vincent Weisser, co-founder and CEO of Prime Intellect, believes companies are looking to move away from a closed-source frontier model, and his company provides the infrastructure to make that transition possible.

“The ability to train AI models shouldn’t just be a few nerds in a glass tower in San Francisco,” he told TechCrunch. “It should be every company, every nation-state.”