Prentis, new AI lab co-founded by Reid Hoffman, Mark Pincus in talks to raise $100M

Hoffman and Pincus Back New AI Agent Venture

Prentis, a new AI research lab co-founded by serial entrepreneur Ritankar Das alongside tech veterans Reid Hoffman and Mark Pincus, is in discussions to raise $100 million at a formidable $1 billion valuation. The lab is focused on developing AI agents capable of observing and automating routine office workflows directly on a computer. This funding effort signals significant investor appetite for specialized models that can execute complex, practical business tasks, such as processing insurance claims or managing customs paperwork, moving beyond the capabilities of general-purpose chatbots.

Technical Edge and Early Traction

Prentis claims its competitive advantage lies in its proprietary, smaller model, which it asserts is more efficient and economical for enterprise deployment. The company has already secured up to $50 million in customer contracts and, according to investor materials, projects an estimated $75 million annualized run rate by the third quarter of this year. Its pitch centers on superior performance and cost-effectiveness compared to larger, more generalized models from industry leaders.

• Model: Hive-32B
• Claimed Performance: Outperforms OpenAI’s GPT-5.4 and Anthropic’s Claude Opus 4.6 on computer-use benchmarks like WindowsAgentArena and ScreenSpot-v2.
• Economic Advantage: Cites a roughly 10 times lower cost per automated task compared to frontier model APIs.
• Team: Recruited over 25 researchers from top labs including OpenAI, Google DeepMind, and Meta.

A Crowded Market for Computer Control

While Prentis is making a strong entry, the market for AI agents that can control computers is becoming increasingly competitive. Incumbents like OpenAI and Anthropic are actively developing their own agentive technologies. Anthropic recently underscored its commitment to this space by acquiring Vercept, a smaller computer-use AI startup. The emergence of heavily-backed players like Prentis indicates a strategic industry race to build the next layer of AI-powered automation, where success will likely be determined by deployment cost and reliability in real-world business environments.

Prentis's billion-dollar valuation talk and substantial early contracts highlight a market shift from foundational model supremacy toward practical, cost-effective AI agents. The key differentiator is no longer just raw capability, but the economic viability of deploying automation across specific, high-value enterprise workflows.