The Daily Byte
Breaking
Welcome to The Daily Byte — real stories from across Pakistan.
Saturday, 21 November 2026 · Pakistan

The Daily Byte

Stories from Pakistan that deserve to be heard
Minor World

Solana Launches Private AI Tools Verified On-Chain

By The Daily Byte ·14 Nov 2026 ·3 min read
Solana logo on a dark background

The Solana Foundation has launched Solana Confidential Compute, an open-source toolkit that lets developers build private AI applications and prove the hardware running them is genuine, with the proof recorded on the Solana blockchain. The launch was announced on 13 November 2026, in a project built with private AI startup Nito, also known as TryNitoAI.

The toolkit is available now in beta on Solana's devnet, the network's testing environment, with a mainnet release planned once a security audit is complete. No date has been given for the audit or the mainnet launch, and the Foundation stresses the release is unaudited and intended for developers building and testing applications for now.

How the verification works

Confidential Compute is built around trusted execution environments, or TEEs: hardware-backed sections of a processor where code and data are isolated from the rest of the machine, including the host operating system and the infrastructure operator. The hardware can produce an attestation, a signed report showing what software is running inside that protected environment.

Solana Confidential Compute moves the checking of that report on-chain. Its verification program checks the attestation and creates a public record that other applications can consult. The AI computation itself happens off-chain, so publishing the verification record does not require publishing the conversation or the private inputs behind it.

The process works in four steps, according to the Solana Foundation's announcement: a machine inside a confidential virtual machine generates a signing key that exists only inside that hardware; the hardware signs a report binding the key to a recent Solana slot; the on-chain program verifies the report and records the key, the hardware type and an expiry date; and the machine then signs its answers with that key, so anyone receiving an answer can check the signature against the on-chain record.

What it supports today

The beta supports hardware attestations from Intel TDX and AMD SEV-SNP. Support for NVIDIA GPU attestations is still in development, which limits the toolkit's reach for now, since much AI work runs on NVIDIA hardware. The tools are open source and free to use.

Nito is the first platform built on the new infrastructure. Its private AI chat is already live, running open-weight models inside attested hardware so prompts can be encrypted on the user's device, decrypted only inside the protected environment, and returned encrypted. Nito is also building a confidential assistant and has opened a waitlist for it. The team behind Nito includes engineers formerly associated with Switchboard.

Why Solana is pushing private AI

The Foundation says the wider opportunity goes beyond chat: personal assistants working with sensitive context, tools processing confidential business information, and trading agents whose strategies must stay private. Other use cases it lists include an AI judge for prediction markets or insurance claims, where a contract pays out only on a ruling signed by a verified machine, and trading agents whose orders execute only if they came from the verified model.

The launch follows Solana Privacy Rings going live on mainnet on 10 November, three days earlier, as part of a broader privacy push that includes a dedicated solana.ai hub for AI products. The core idea, as Crypto Briefing puts it, is to replace trust with proof: rather than asking users to believe an AI provider is handling their data properly, an app can point to an on-chain attestation showing what hardware and code are in use. One caveat remains: TEE-based privacy still rests on trusting the chip makers whose hardware issues the attestations, so the blockchain verifies the receipt but not the hardware itself.