Prentis AI Lab Seeks $100M at $1B Valuation for Computer Control

Prentis targets $1 billion valuation in new funding round

Prentis, an AI research lab developing models that can control computers to automate office workflows, is in discussions to raise $100 million at a $1 billion valuation, according to two sources with knowledge of the talks.

The startup was co-founded in April by Ritankar Das, LinkedIn co-founder Reid Hoffman, and Zynga founder Marc Pincus. Prentis trains AI models by observing how office workers navigate routine tasks across documents and systems, then builds agents capable of executing those workflows autonomously.

The company has already secured contracts valued at up to $50 million with customers including a healthcare management service organization, a manufacturer, and clothing producers. According to investor materials, Prentis projects a $75 million annualized run rate by the third quarter of this year, though the company notes these figures represent estimated annualized value based on contracted fees equal to 20% of realized savings—not recognized revenue—and remain performance-dependent.

Technical claims and market positioning

Prentis asserts its Hive-32B model outperforms competing systems from OpenAI and Anthropic on computer-use benchmarks. The company cites superior performance on WindowsAgentArena, which measures end-to-end task completion on real Windows applications, and ScreenSpot-v2, which evaluates a model's ability to identify correct on-screen controls.

The startup's pitch emphasizes cost efficiency, claiming roughly 10 times lower cost per task than frontier APIs by running a smaller, more economical model. These benchmark results have not been independently verified.

Prentis envisions building custom agents for specific business needs—handling insurance claims, automating customs duty refund exceptions, and similar tasks that currently require humans to locate and process paperwork across multiple systems.

Crowded competitive landscape

The computer-use AI segment has attracted significant attention from major players. Anthropic, OpenAI, and Mira Murati's Thinking Machines are all developing competing computer-control agents. Anthropic acquired Seattle-based computer-use startup Vercept earlier this year, absorbing its founders and discontinuing its product.

Prentis has assembled a team of more than 25 employees, including researchers formerly at OpenAI, Google DeepMind, Meta, Tencent, and Alibaba.

Why it matters

Computer-use AI represents a fundamental shift from chatbots and coding assistants toward agents that can execute complex, multi-step business processes autonomously. If Prentis can deliver on its cost and performance claims, the technology could unlock automation for routine office work that has remained stubbornly manual despite decades of software tools. The $50 million in early contracts suggests enterprise appetite for this capability is real, though the performance-based revenue model also signals customers remain cautious about unproven AI agents handling critical workflows.

Background on the founders

Das, now 31, serves as CEO and founded Titan in 2014—a holding company that builds and operates AI companies. He was UC Berkeley's youngest University Medalist in over a century, graduating at 18 before earning a master's at Oxford. Titan has launched multiple healthcare AI ventures, including Tala Health, which raised $100 million last year, and Forta Health, which secured $55 million in 2024.

For Hoffman and Pincus, Prentis represents a side venture. Hoffman recently stepped down from Microsoft's board to focus on Manas AI, a drug-discovery startup. Pincus now runs investment firm Reinvent Capital, with Hoffman serving as senior adviser.

These details were first reported by TechCrunch. Prentis did not respond to requests for comment.