# Solana Foundation launches Confidential Compute with TryNitoAI

> Source: <https://cryptobriefing.com/solana-foundation-confidential-compute-trynitoai/>
> Published: 2026-11-13 17:17:23+00:00

# Solana Foundation launches Confidential Compute with TryNitoAI

solana coin

An open-source toolkit for verifying private hardware on-chain arrives on devnet as an unaudited beta aimed at private AI and trading agents

The Solana Foundation wants AI agents to keep secrets, and it wants the blockchain to prove they are keeping them.

On November 13, 2026, the foundation and private AI startup TryNitoAI, which goes by Nito, launched [Solana](https://cryptobriefing.com/markets/solana/) Confidential Compute. It is an open-source set of tools for building private AI and trading agents, and it currently runs as an unaudited beta on Solana’s devnet.

That last detail matters. Devnet is the network’s practice field, so nothing here is handling real money yet.

## What Solana Confidential Compute actually does

The launch has two main pieces: an open-source SDK and an on-chain verification program. Together, they let developers check hardware attestations from Trusted Execution Environments, or TEEs, directly on Solana.

A TEE is a walled-off section of a processor where code can run without the rest of the machine peeking in. An attestation is the hardware’s signed statement about what is running inside that walled-off section.

Normally, checking that receipt happens off to the side, out of public view. Solana’s new program moves the verification on-chain, where anyone can see whether a given piece of hardware passed the check.

For now, the system supports two flavors of attestation: Intel’s TDX and [AMD](https://cryptobriefing.com/markets/amd/)’s SEV-SNP. Support for [NVIDIA](https://cryptobriefing.com/markets/nvidia/) GPU attestations is still in development.

The tools are being offered openly and at no cost. Because of that, any developer can take part in verifying hardware attestation records on the network, rather than leaving the job to a single gatekeeper.

### AI, tech, and the markets they move—in one daily briefing.

Daily. Free. Join 34,000+ readers across crypto, finance, and policy.

## Nito goes first

Nito is the first platform built on the new infrastructure. The company runs open-weight AI models inside attested hardware, which allows for end-to-end encrypted interactions with those models.

Nito’s private AI chat is live at trynito.ai. The company is also building a confidential assistant and has opened a waitlist for it.

The team behind Nito is made up of engineers formerly associated with Switchboard.

The Solana Foundation has laid out several use cases it sees for the technology:

- Private AI applications
- Personal assistants
- Agents that handle sensitive business information
- Private trading strategies

## The road to mainnet

A mainnet release is expected after a security audit is completed. No timeline has been given for either the audit or the launch.

## Part of a broader AI push

Confidential Compute fits into Solana’s larger ambition to serve as the base layer for AI applications. That effort includes a dedicated solana.ai hub and a string of privacy-focused releases.

The most recent came just days earlier. Solana Privacy Rings went live on mainnet on November 10, 2026, three days before Confidential Compute arrived on devnet.

## What this means for developers and the network

The core idea is to replace trust with proof. Rather than asking users to believe an AI provider is handling their data properly, an app built on these tools can point to an on-chain attestation showing what hardware and code are in use.

The software is unaudited, it runs only on devnet, and there is no date for mainnet. Hardware coverage is also incomplete: Intel and AMD attestations work today, but NVIDIA GPU support is still being built.

There is also a structural point worth noting. TEE-based privacy rests on trusting the chip makers whose hardware generates the attestations, so the blockchain verifies the receipt but still relies on the hardware that issued it.

**Disclosure:** This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our [Editorial Policy](https://cryptobriefing.com/editorial-policy/).
