ZeroDrift raises $10 million to protect AI models from themselves

1 day ago 6
A well-dressed antheral   leans against a neutral colored wallImage Credits:ZeroDrift

5:32 AM PDT · June 2, 2026

As enterprises troubleshoot their AI systems, governance has emerged arsenic a cardinal challenge. Some are taking a dual approach: One exemplary to grip incoming queries, portion a 2nd exemplary keeps the archetypal 1 from getting into trouble.

That’s the premise of ZeroDrift, a caller AI compliance work that announced a $10 cardinal effect circular connected Tuesday. (Investors see a16z Speedrun, Reign Ventures, PitchDrive Ventures, and U&I Ventures, among others.) The institution deals wholly with the 2nd portion of the system, sitting betwixt AI models and extremity users to emblem and regenerate immoderate messages that mightiness contiguous a compliance problem.

It mightiness look unusual to physique an AI strategy to close different AI systems’ mistakes — but ZeroDrift’s correction strategy has a fewer architectural advantages implicit the models it volition beryllium correcting. The strategy is triggered by accepted programs that deterministically use known compliance standards similar SOC 2 oregon GDPR. The LLM lone comes successful erstwhile a connection has been flagged, rewriting a compliant mentation of the aforesaid message.

“We’re capable to place deterministically, what are each the regulated areas, what’s the usurpation that’s being broken, and past we person LLMs that tin bash the rewrites,” Aroomoogan says.

Critically, the full strategy tin beryllium tally with little latency and much reliability than a accepted LLM. It’s the company’s superior vantage implicit large labs similar OpenAI and Anthropic, which are often already contiguous successful the underlying system.

The astir evident usage lawsuit is for AI chatbots, which are already deployed successful beforehand of consumers wherever determination tin beryllium superior consequences for rogue answers. But Aroomoogan sees the full addressable marketplace arsenic overmuch larger, perchance spanning AI-generated messages that quality beings ne'er see, generated lone wrong automated systems. So far, it’s a comparatively tiny marketplace — but it’s 1 that volition turn arsenic AI proliferates.

If the fundraise is immoderate indication, there’s a batch of pent-up request for that product. “It was astir apt the fastest fundraising I’ve done successful my life,” CEO Kumesh Aroomoogan says, crediting Andressen Horowitz’s assistance structuring the effect round. “We closed wrong 3 weeks, and we volition beryllium oversubscribed by 3x connected the amount.”

When you acquisition done links successful our articles, we whitethorn gain a tiny commission. This doesn’t impact our editorial independence.

Russell Brandom has been covering the tech manufacture since 2012, with a absorption connected level argumentation and emerging technologies. He antecedently worked astatine The Verge and Rest of World, and has written for Wired, The Awl and MIT’s Technology Review. He tin beryllium reached astatine russell.brandom@techcrunch.com oregon connected Signal astatine 412-401-5489.

Read Entire Article