Prime Intellect Raises $130 Million to Help Enterprises Build Their Own AI Agents
AI infrastructure startup Prime Intellect secured $130 million in Series A funding, reaching a $1 billion valuation, to give enterprises the tools and compute power to build independent AI agents.
Prime Intellect, a startup offering decentralized computing and software for AI development, has raised $130 million in a Series A round. The deal values the company at $1 billion, just one year after its founding. Radical Ventures led the investment, with participation from Nvidia Ventures, Intel Capital, Dell Technologies Capital, and Iconiq. Perplexity founder Aravind Srinivas and Box CEO Aaron Levie also joined the round as individual investors.
The company's pitch is straightforward: building high-performing AI agents should not require the specialized talent and massive compute budgets available only to frontier labs like OpenAI or Anthropic. Prime Intellect offers a modular, full-stack platform that includes a reinforcement learning framework, evaluation tools, and access to the required hardware. Enterprises can use it to train models on their own tasks without sending proprietary data to third-party providers.
Demand for that kind of independence is growing. A number of enterprises are rethinking their reliance on closed-source AI providers, citing data privacy concerns and unpredictable service decisions. Prime Intellect pointed to Anthropic's sudden shutdown of Fable as an example of the risks companies face when they have no control over their AI infrastructure. The startup already counts fintech firm Ramp and automation platform Zapier among its customers and has reached an annualized revenue run rate of $100 million.
The reinforcement learning approach allows organizations to train agents that often outperform general-purpose models on specific tasks. Ramp, for instance, used the platform to build a specialized agent for spreadsheet analysis. Prime Intellect says the resulting system beat frontier models on accuracy while running faster and at lower cost. CEO Vincent Weisser has been direct about the company's broader ambition: the ability to train AI should not be limited to a small group of developers in San Francisco but should be accessible to enterprises and nation states everywhere.
The Africa Angle
For African startups and government agencies, the emergence of sovereign AI infrastructure carries real practical weight. API costs from global frontier labs are priced in dollars, creating pressure on teams operating in naira, cedis, or shillings. Data privacy protections in many African jurisdictions also sit uncomfortably with arrangements that require sending sensitive local data to servers abroad. A platform that enables local training and keeps data within an organization's control addresses both problems. It also opens the door to building agents trained on African languages and market-specific data, products that general-purpose Western models consistently underserve. The sudden shutdown of services like Fable is a concrete reminder that dependency on external providers carries strategic risk. Infrastructure that reduces that dependency is not just a technical preference; for many African organizations, it is a policy priority.
The power to build intelligence is worth little if it stays concentrated in a handful of cities on one continent.
