I’m joining Anthropic. Here’s the official OPAQUE post: From Berkeley Research to the Future of AI: OPAQUE’s Next Chapter. And here’s a Linkedin Post on the topic. This is the personal version: what we built, who we built it with, and the people I’d hire tomorrow if I were you.
Three years ago
I joined OPAQUE in October 2023. The founders, Ion Stoica, Raluca Ada Popa, Rishabh Poddar and Chester Leung, invited me to join them on their journey to make the world a little safer. They had spent years at UC Berkeley’s RISELab proving that data can stay protected while it’s being used. My job was to turn that research into a company and a market.
In my first conversations with analysts and journalists I could barely articulate what we did. I knew what we needed to achieve. I wasn’t sure it was possible. People I trust and respect told me it was not. It was and we proved this.
We defined a category
We called it confidential AI: verifiability for the AI stack. Cryptographically enforced proof, signed by hardware, that governance and security policies hold while AI runs on sensitive data, instead of a promise that they will. It doesn’t cover every policy or close every hole. It covers many of the critical ones, and it replaces “trust us” with evidence.
That work now secures AI systems at government agencies, national sovereign AI projects, enterprise software vendors and regulated industries.
We didn’t build the category alone. We created the Confidential Computing Summit, and from day one it was about building a security and governance layer for AI, in the open. We were privileged to have the sponsorship and participation of NVIDIA, Intel, AMD, Microsoft, Google, Apple, ServiceNow, Google DeepMind, Anthropic, AWS and many others. This year we were thrilled to have the Linux Foundation take over operations of CCSummit; their world class team will continue to make this event the annual event for AI security and governance.
Then Apple shipped Private Cloud Compute and made the architecture the standard for running AI on personal data. Meta, Google and OpenAI followed with their own versions. A research idea pioneered by OPAQUE’s founders became the way the largest technology companies on earth run AI on the data that matters most.
This year
We secured many terrific customer and partners across enteprise independent software vendors, regulated industries, governmental agencies, global systems integrators, and AI labs. Earlier this year, in January, we closed our Series B at a $300 million valuation.
We partnered with the brilliant team at the Technology Innovation Institute in Abu Dhabi — Dr. Najwa Aaraj, Dr. Victor Mateu and Dr. Victor Sucasas — and with the UAE Cyber Security Council to create a governance and security framework for verifiable AI systems. It’s becoming a national standard, and it’s shaping how the region and the rest of the world think about governing AI.
In June we launched AgenTrust, an open source community for verifiable agent governance. It took off. Nearly a million downloads in just over three months. I’ve since learned that several companies have launched on that code, and large enterprises and government agencies, including in regulated industries, have adopted it as core components of their AI stack.
We were also among the first to argue that AI agents have to be treated as insiders, with cryptographic enforcement behind the policies that govern them. This past weekend Satya Nadella, Chairman and CEO of Microsoft, made the same case.
Why Anthropic
It’s been an incredible run, and I’m thrilled to be joining Anthropic. Their culture, their mission and their values align with mine, personally and professionally, and I’m thrilled to join a journey and a mission that matters to the world.
I’m grateful to our customers, who trusted us with their most sensitive workloads before the category had a name. To our partners, who built it with us. To our investors and board. And to the OPAQUElings, who worked incredibly hard to make the world a little better and a little safer.
Hire these people
Not every OPAQUEling is joining Anthropic. Some of the most talented people I’ve worked with are looking for their next role right now, in engineering, design and go-to-market. If you’re hiring, find them on LinkedIn or reach out to me directly. Every one of them will be an accelerant to your mission and your business.
We set out to make the world a little better and a little safer. I’m proud to say that we did, with the help of our customers, our partners and many others. Thank you all.
Here are the incredibly talented and skilled people who made OPAQUE a success. Some are actively pursuing their next gig. Feel free to contact me directly for more information on the OPAQUE talent. 🙂
