Confidential AI

Conf AI makes it trivial to deploy AI workloads inside TEEs. e2e private inference, training, and agents.

[Confidential AI @Confi_AI]

@hyperbolic_labs Do you have any B200s or B300s available that we can set up in CC mode?

[Confidential AI @Confi_AI]

AI should be personal. And personal means private.

Sign up for the beta here 👇

https://t.co/ErgljLdn9t

Launch post + API docs: https://t.co/VwQjwYE3cU

[Confidential AI @Confi_AI]

Teams in finance, healthcare, and working on privacy-sensitive tasks with AI get stuck with clunky on-prem workarounds.

Why? Because the only thing protecting their data in a cloud is a legal agreement. That’s not enough.

We're fixing that today with Confidential Agents. 🧵

[Confidential AI @Confi_AI]

Don’t take our word for it. Verify it yourself.

All Confidential Agents come with an open-source CLI that verifies that no one has access to your host, that your inference is private, and that everything runs in TEEs.

[Confidential AI @Confi_AI]

We are now a proud member of the @linuxfoundation and the @ConfidentialC2.

It's cool to see our logo on their members pages.

[Confidential AI @Confi_AI]

This is what actually-private AI inference looks like.

Prompt encrypted on your device, decrypted only inside a hardware-attested enclave. Cloud, hypervisor, host OS, vendor: all see ciphertext.

Two places hold plaintext: your device and the enclave. Everything else is locked out by hardware.

[Confidential AI @Confi_AI]

[███]
New logo. A redacted bar inside square brackets.

Most security logos are padlocks and shields. They all say "trust us." Confidential Computing is the opposite. The contents are not protected by us. They are unreachable.

Confidential computing in a few characters.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

New paper from the team at @Confi_AI: Kettle, Attested Builds.

Run the build inside a TEE. Staple a hardware-signed attestation of the pipeline, inputs, and outputs to the artifact.

Provenance down to the git commit.

https://t.co/PwrzwnwhSU

[Confidential AI @Confi_AI]

Accepted into NVIDIA Inception.

Confidential runs AI inside TEEs. Private inference. Private weights. Private training. Verifiable end to end.

Now with closer ties to the GPU stack underneath it all.

#NVIDIAInception

[Confidential AI @Confi_AI]

Confidential computing for Kubernetes, without rethinking the control plane.

Our C8s whitepaper is out.

[geomatrick 🌐 👷 💻 @patrickwoodhead]

AI workloads on shared infrastructure have a three-body problem.

Artifact owners, compute providers, and end users all need confidentiality from each other.

Today my team at @Confi_AI published C8s, a confidential Kubernetes architecture that solves it.

[Confidential AI @Confi_AI]

PrivateClaw, the world's first e2e private agent backed by hardware encryption, launched today on Producthunt!

PrivateClaw is powered by the @Confi_AI stack.

https://t.co/ucI5PN21oh

[Confidential AI @Confi_AI]

By 2029, Gartner predicts more than 75% of operations processed in untrusted infrastructure will be secured in-use by confidential computing.

Only 75%?

We're building for 100%.

[Confidential AI @Confi_AI]

Welcome to Confidential, @AnomalRoil

[ansgar grunseid @ansgargg]

✨ im thrilled to share that the swiss ritter of randomness, @AnomalRoil, joins our merry band of misfits at @Confi_AI 🎉

he and batman have never been seen in the same room at the same time. just sayin'.

[Confidential AI @Confi_AI]

We are in SF this week around HumanX. If you're thinking about private inference, secure model deployment, or confidential computing, let's grab a coffee.

[Confidential AI @Confi_AI]

Trusted Execution Environments, explained.

A useful analogy is the transition from HTTP to HTTPS.

With HTTP, you sent your data in plaintext but couldn't confirm who you were talking to.

HTTPS allows you to confirm who you're talking to and encrypt your data in transit.

But once your data arrives at the server, it's decrypted and processed in plaintext. This means the system administrator can see your data.

TEEs go a step further. Your data stays encrypted by hardware during computation. You can verify who you're talking to, what software they're running, and that your data is private throughout.

The sys admin can't access the data. No one can.

All of this, like HTTPS, adds a negligible performance cost.

[Confidential AI @Confi_AI]

Confidential AI is the most important infrastructure problem of today.

Exposing our data and identity to the tech giants was bad. The same is happening with the frontier labs in fast forward mode. As AI infiltrates every workflow, confidentiality will necessarily become table stakes.

Yet the labs want this too. They want to protect their model weights. Privacy is essential from all perspectives.

Confidential computing is hitting an inflection point. It is crossing the chasm from theory and research into production. But this problem is not yet solved.

That's why today we decided to rename ourselves Confidential AI.

This new name describes exactly what gets the team out of bed every day.

Lunal as a name has worked for us until now: an engineering team solely focused on shipping by daylight or moonlight. Until a few weeks ago, our domain https://t.co/E8wHxXZyHU just pointed to our Github page. A signal that we cut out the fluff and focus on the load-bearing work.

The name has changed but the approach hasn't. Same team, same roadmap, same hellbent desire to ship high-quality, privacy-preserving products. We've shipped confidential inference, training, fine-tuning, and agentic flows. And now, finally, we got round to shipping a new name and a website.

We are Confidential AI and we do Confidential AI.

Find us at https://t.co/QQte4eeoQj. If you're building with AI and care about who sees your data or model weights, we should talk.

[Confidential AI @Confi_AI]

It was a great evening 🙌

[Initialized Capital @Initialized]

4 founders on what it means to work at a startup right now.

Why right now matters:
"We melted sand, turned it into chips, and the chips are thinking. That is an enormous step forward — and to not be taking advantage of that would be a disservice to your career, and maybe a disservice to progress for humanity." — @lucas0choa, @automat_ai

[Confidential AI @Confi_AI]

Huge thanks to @Initialized for having Lunal's CEO @ansgargg on stage last night at the 3rd annual Initialized Talent Summit. We really enjoyed the event.

[Confidential AI @Confi_AI]

https://t.co/ZzbjpDI9ZS

[Confidential AI @Confi_AI]

We built a simulator for c8s, our confidential Kubernetes product with autoscaling.