Three months after launching in April, Prentis AI lab is in talks to raise $100 million at a $1 billion valuation, with co-founders Ritankar Das, Reid Hoffman and Marc Pincus betting that AI agents that operate computers can become bigger than coding, according to TechCrunch.
The talks are not closed. TechCrunch cited two people familiar with the discussions, which means the amount, valuation, investors and timing can still shift. But the ambition is already clear: Prentis wants to train computer-use models that watch how office workers move through documents and systems, then automate those workflows.
Three months after launch, Prentis AI lab is pitching a $1 billion valuation
Prentis is building AI agents that can control computers to handle routine work, rather than simply answer prompts or suggest code. The company’s pitch centers on tasks such as handling insurance claims and automating customs duty refund exceptions without requiring a human to track down paperwork.
That matters because Prentis is aiming at the dullest parts of enterprise work. Those are also the parts companies often want to reduce, if the software can do the job reliably.
The company has already signed contracts worth up to $50 million with several customers, according to TechCrunch’s sources. 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. That figure comes with a major caveat.
Prentis’s pitch deck says 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 caveat is doing real work. It means the headline commercial numbers depend on execution and realized savings, not booked revenue already sitting on the income statement.
April launch put office workflows at the center of the Prentis pitch
Prentis was launched in April and is training models to learn how office workers navigate routine workflows across documents and systems. The product thesis is that the next major AI use case may not be coding. It may be automating the repeated computer actions that sit inside insurance, healthcare, manufacturing and trade processes.
The company calls its model Hive-32B. In its own materials, Prentis says Hive-32B 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 whether a model can locate the correct on-screen control.
TechCrunch has not independently verified those benchmark results.
| Prentis claim | Source basis | Caveat |
|---|---|---|
| Hive-32B beats GPT-5.4 and Claude Opus 4.6 | Prentis pitch materials cited by TechCrunch | Not independently verified by TechCrunch |
| Roughly 10 times lower cost per task | Company claim in pitch deck | Based on Prentis’s own comparison with frontier APIs |
| Contracts worth up to $50 million | Two people familiar with the discussions | Figures are tied to performance and final execution |
| Estimated $75 million annualized run rate by Q3 | Investor materials obtained by TechCrunch | Not recognized revenue |
XOOMAR analysis: Prentis’s clearest strategic bet is cost. If a smaller model can complete real software tasks at meaningfully lower cost, the deployment argument changes. The harder question is whether those agents can keep working outside controlled tests, where documents, interfaces and business exceptions rarely behave cleanly.
Benchmark claims and signed contracts are carrying the funding story
The fundraising case appears to rest on three pillars: notable founders, early customer contracts and claimed technical efficiency. For a company only launched in April, that is a lot to put in front of investors quickly.
Prentis says its advantage comes from running a smaller, cheaper model. The company claims roughly 10 times lower cost per task than frontier APIs, arguing that this makes it more economical for everyday workflows.
That is a sharper claim than “better AI.” It says Prentis can be cheaper where repetitive work happens at volume.
The field is crowded. TechCrunch reported that Anthropic, OpenAI and Mira Murati’s Thinking Machines are also working on AI agents for computer use, citing one of its sources. Anthropic has also been buying talent directly in the category: it acquired Seattle computer-use startup Vercept earlier this year, folded in its founders and shut down the product.
For readers following XOOMAR’s separate AI coverage, see $4.5M Imagi Bet Pulls Vibe Coding Into K-12 Schools and Stripe OpenRouter Talks Put $10B AI Tollbooth in Play. Those are separate stories, not evidence for the Prentis round, but they sit near the same investor debate over where AI turns from impressive demos into repeatable usage.
Hoffman, Pincus and Das give the round founder-heavy signaling
The co-founder list is part of the story. Reid Hoffman co-founded LinkedIn, is a Greylock partner, was an early OpenAI investor and co-founded Inflection AI with Mustafa Suleyman before Microsoft absorbed most of that team in 2024.
Hoffman 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.
Marc Pincus, the Zynga founder, now runs investment firm Reinvent Capital, with Hoffman as a senior adviser. He also published a memoir, “Life at the Speed of Play,” last month.
Das brings the operating company angle. He is CEO of Prentis and founder of Titan, a holding company that builds and operates AI companies. TechCrunch notes that Titan-backed or Titan-founded businesses include Tala Health, Forta Health and Dascena, which was acquired by CirrusDx in 2022.
Prentis has hired more than 25 employees, including researchers who previously worked at OpenAI, Google DeepMind, Meta, Tencent and Alibaba, according to its website.
The next decision point: capital first, dependable agents after that
The immediate question is whether the Prentis AI lab round closes near the reported $100 million target and $1 billion valuation. The next question is harder: how much of the pitch survives contact with customer workflows.
The company did not respond to TechCrunch’s request for comment. That leaves investors and customers working from reported fundraising talks, pitch deck claims, early contract figures and Prentis’s own benchmark assertions.
XOOMAR analysis: the risk is not that office automation lacks a market. The risk is that computer-use agents must be dependable in messy, high-friction workflows where a bad click or missed document can break the process. Prentis is chasing one of AI’s most valuable promises, but the next proof point won’t be a benchmark slide. It will be whether its agents can complete real work with minimal human correction, at the cost Prentis says it can deliver.
The Bottom Line
- Prentis is seeking a $1 billion valuation just three months after launch, signaling intense investor interest in AI agent startups.
- The company is targeting enterprise workflows like insurance claims and customs refund exceptions, where automation could cut major back-office costs.
- Its reported revenue projections depend on performance-based savings, making execution and customer results central to whether the valuation holds.
Originally published on XOOMAR. For more news and analysis, visit XOOMAR.
Top comments (0)