Prentis is entering one of the most watched areas in artificial intelligence: models that can operate computers on behalf of people. The new AI research lab, launched in April, is focused on computer use models designed to learn how office workers move through routine workflows across documents and systems.
According to two people familiar with the discussions cited by TechCrunch, Prentis is in talks to raise $100 million at a $1 billion valuation. The company was co-founded by serial entrepreneur Ritankar Das along with Reid Hoffman and Mark Pincus, bringing together an AI operating-company background with well-known technology investors and founders.
A bet on routine office workflows
The core idea behind Prentis is straightforward: many office jobs involve repeated movement between documents, internal systems, screens, and approvals. Prentis is training models to understand those patterns and eventually control computers to automate the work.
The company is targeting workflows where AI agents could reduce the need for a person to manually gather information or hunt through paperwork. TechCrunch reports examples such as handling insurance claims and automating customs duty refund exceptions.
That positioning makes Prentis different from AI tools that only generate text, summarize documents, or assist with coding. Its focus is on computer use AI: systems that can identify on-screen controls, navigate real applications, and complete tasks from end to end.
For customers, the appeal is not just that an AI agent can answer a question. It is that the agent could potentially take action across existing systems, especially in repetitive operational processes where small delays can accumulate.
Early contracts and performance-based claims
Prentis has already signed contracts worth up to $50 million with several customers, according to the two people familiar with the discussions cited by TechCrunch. Those customers include a healthcare management service organization, a manufacturer, and goods and clothing manufacturers.
Investor materials obtained by TechCrunch also predict an estimated $75 million annualized run rate by the third quarter of this year. The source article notes an important caveat from Prentis' pitch deck: those figures reflect estimated annualized value based on a contracted fee equal to 20% of savings realized, not recognized revenue, and are "performance-dependent and subject to final execution."
That distinction matters. A contract tied to savings is not the same as revenue already booked. It suggests that Prentis' business model, at least as described in the pitch materials, depends on whether its agents can deliver measurable results in customer workflows.
Still, the reported customer activity gives the company a concrete commercial story at an early stage. It is not only presenting a research vision; it is pitching computer use models as a practical enterprise automation product.
What Prentis says its model can do
By its own account, Prentis says its Hive-32B model outperforms rivals, including OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6, on two computer-use benchmarks.
The first benchmark named in the source is WindowsAgentArena, which measures end-to-end task completion on real Windows applications. The second is ScreenSpot-v2, which tests whether a model can locate the correct on-screen control.
Those benchmarks line up closely with the company's stated goal. If an AI agent is going to automate office workflows, it needs to do more than understand language. It must recognize interfaces, choose the right controls, and complete tasks through software that workers already use.
Prentis also argues in its pitch deck that its advantage comes from using a much smaller, cheaper model. The company claims roughly 10 times lower cost per task than frontier APIs, saying that makes its approach more economical for everyday workflows.
TechCrunch says it has not independently verified the company's benchmark results. That means the claims are part of Prentis' pitch, not confirmed findings from the publication.
A crowded computer use AI race
Prentis is betting that automating everyday office tasks will soon outpace coding as AI's biggest use case. That is a large claim, and the source makes clear that the market is already crowded.
Anthropic, Open AI, and Mira Murati's Thinking Machines Lab are also working on AI agents for computer use, one of the sources told TechCrunch. Anthropic has also been acquiring talent in the category directly. It bought the Seattle computer-use startup Vercept earlier this year, folding in its founders and shutting down its product.
That competitive backdrop raises the stakes for Prentis. If computer use AI becomes a central enterprise category, the winners will likely need strong models, low operating costs, customer-specific agents, and enough reliability to handle real workflows without constant human correction.
Prentis' pitch appears to address those points directly: smaller model economics, benchmarks tied to real applications and on-screen controls, and customer contracts connected to operational savings. But the company has not commented publicly in response to TechCrunch's request.
The founders behind the lab
Ritankar Das, CEO of Prentis, is also the founder of Titan, a holding company that builds and operates AI companies. Das, now 31, was UC Berkeley's youngest University Medalist in more than a century, graduating at 18 with a double major in bioengineering and chemical biology before earning a master's in biomedical engineering at Oxford.
He founded Titan in 2014 after dropping out of an AI PhD program at Cambridge, where he'd been a Gates Cambridge Scholar. Das has described Titan as an intentional throwback to an old-fashioned holding-company model like Berkshire Hathaway, funded by its own exits rather than outside limited partners.
Other Titan-launched businesses include AI-powered virtual care provider Tala Health, which raised a $100 million seed round last year, and Forta Health, an autism care startup that raised $55 million led by Insight Partners in 2024. Titan-founded disease prediction company Dascena was acquired by CirrusDx in 2022.
Prentis is described as more of a side project for its other two co-founders. Hoffman, the LinkedIn co-founder and Greylock partner, said last month that he was stepping down from Microsoft's board after nearly a decade to go "founder mode" on Manas AI, an AI drug-discovery startup he is also backing. He was an early OpenAI investor and co-founded Inflection AI with Mustafa Suleyman before Microsoft absorbed most of that team in 2024.
Pincus, the Zynga founder, now runs the investment firm Reinvent Capital with Hoffman as a senior adviser, and published a memoir, "Life at the Speed of Play," last month. Prentis has already hired more than 25 employees, including researchers who previously worked at OpenAI, Google DeepMind, Meta, Tencent, and Alibaba, according to its website.