Prentis, co-founded by Reid Hoffman and Mark Pincus, in talks to raise $100M for AI agent lab
Startup developing computer-use models for office task automation seeks $1B valuation amid $75M annualized run rate projections.
1 source · cross-referenced
- Prentis, a new AI lab co-founded by Reid Hoffman and Mark Pincus, is in funding talks to raise $100M at a $1B valuation.
- The lab is developing AI agents to automate routine office tasks, with contracts worth up to $50M already signed.
- Prentis claims its Hive-32B model outperforms rivals on computer-use benchmarks but has not been independently verified.
- The startup projects a $75M annualized run rate by Q3 2026 based on contracted fees tied to realized savings.
Prentis, a newly formed AI research lab focused on computer-use models, is in talks to raise $100 million at a $1 billion valuation, according to two people familiar with the discussions. The lab was launched in April and is co-founded by serial entrepreneur Ritankar Das alongside tech figures Reid Hoffman and Mark Pincus. Prentis is developing AI agents designed to automate routine office tasks by learning how workers navigate workflows across documents and systems. Potential applications include handling insurance claims and automating customs duty refund exceptions without manual paperwork retrieval.
The startup has already secured contracts worth up to $50 million with multiple customers, including a healthcare management service organization, a manufacturer, and goods and clothing manufacturers. Prentis projects a $75 million annualized run rate by the third quarter of 2026, citing investor materials. These figures are based on contracted fees equal to 20% of savings realized by customers, not recognized revenue, and are described as performance-dependent and subject to final execution.
Prentis asserts that its Hive-32B model outperforms rivals, including OpenAI’s GPT-5.4 and Anthropic’s Claude Opus 4.6, on two computer-use benchmarks: WindowsAgentArena, which measures end-to-end task completion on real Windows applications, and ScreenSpot-v2, which tests a model’s ability to locate the correct on-screen control. The company attributes its performance to running a smaller, more cost-effective model, claiming roughly 10 times lower cost per task than frontier APIs. TechCrunch has not independently verified these benchmark results.
The lab’s focus on automating everyday office tasks reflects a broader industry bet that agentic automation will surpass coding as AI’s primary use case. However, the space is already crowded, with competitors like Anthropic, OpenAI, and Mira Murati’s Thinking Machines Lab also developing AI agents for computer use. Anthropic has further signaled its commitment to the category by acquiring the Seattle-based computer-use startup Vercept earlier in 2026, integrating its founders and discontinuing its product.
- Jul 24, 2026 · The Verge — AI
Meta AI assistant update adds calendar integration and daily briefings
Trust78 - Jul 24, 2026 · Latent Space — swyx
Black Forest Labs releases FLUX 3 Video and FLUX-mimic, unifying multimodal generation and robotics control
Trust69 - Jul 24, 2026 · arXiv cs.AI
Researchers propose AINTMA, a multi-agent AI system for autonomous software test management with generative quality analytics
Trust79