Why I Joined Lunal as Head of Product

AI is the most powerful technology of our lifetimes, and right now it's being built in a way that concentrates power, data and trust in a handful of companies. I've spent my career pushing against exactly that. Lunal is the most compelling antidote I've found.

We now use LLMs and agents for everything - an oracle for all information, a team of workers at your side all day. Yet every time I press send on a prompt, a gremlin in my head raises a red flag: are you sure you want to share this with that frontier lab?

Across all my prompts, Anthropic and OpenAI probably know me better than I know myself. That already makes me uneasy - and it's not just a personal paranoia. The US Department of Defense recently pressured Anthropic to make its technology available for surveillance and penalised them when they refused. The infrastructure we're all using every day is already a geopolitical battleground.

The same problem hits companies far harder. They want the productivity gains of LLMs but won't send proprietary data into the system. They need private computation, often in a specific geography or on premises.

Then there are the AI labs themselves. Their model weights are among the most valuable assets a company, and arguably a country, currently holds. As they distribute their models across more hardware and infrastructure partners, they need hard guarantees those weights stay secure.

And those partners want exactly the same thing - confident users and confident labs means more business.

Three parties, one shared need: end-to-end AI privacy.

Trusted Execution Environments (TEEs) are the most compelling and closest-to-market solution. All parties get end-to-end verifiability that data is encrypted in transit and only decrypted inside a computing enclave that no one can access - not even the person at the hyperscaler who deployed it.

The first AI lab to offer private inference as a service tier will trigger a domino effect. Everyone else will have to follow.

The catch: migrating AI deployments onto TEEs is hard. That is exactly what Lunal exists to solve. Learn more at Lunal.dev.

Lunal raised a seed round at the end of 2025 and is growing fast. I'm thrilled to be working with a team that has the depth to pull this off.