In This Article
The Solana Foundation has introduced Solana Confidential Compute — an open-source toolkit built on Nito that enables developers to create private artificial intelligence applications and verify their computing hardware on Solana.
The Solana Foundation’s November 13 announcement describes a software development kit and an on-chain verification program developed by Nito in collaboration with the Solana Foundation. Nito integrates this infrastructure into its private artificial intelligence platform, while other teams can use the same open-source components.
“Private AI, verified on-chain,” the Solana Foundation wrote on X, presenting the tools for applications that process personal conversations, confidential business information and private trading strategies.
The beta version supports hardware attestation from Intel TDX and AMD SEV-SNP, while attestation for NVIDIA graphics processors is still under development. In the same thread, the Foundation described the release as unaudited and intended for further development. The announcement states that a mainnet version will be released soon, but no release date is specified.
Solana Confidential Compute verifies hardware
According to the Foundation, confidential computing protects information while it is being processed by a machine. The system uses trusted execution environments (TEEs), which isolate code and data from the surrounding infrastructure, including the host operating system and its operator.
Inside a confidential virtual machine, the software generates a signing key that remains within the protected hardware. A hardware-signed report links the key to the latest Solana slot, and the verification program checks the report before recording the key, hardware type and expiration date.
When the machine produces a response, it signs it with that key. The application receiving the response can verify its signature and consult the corresponding record on Solana before accepting it.
The Foundation states that AI computations take place off-chain. Publishing a verification record does not require publishing the user’s correspondence or personal data.
In private chat, requests are encrypted before they leave the browser
The Solana chat template provides an example for developers using Next.js and React. According to the documentation, the browser verifies the confidentiality of the environment before encrypting the message with a key. A failed check prevents the request from being sent.
The encrypted messages pass through the application server and Nito into a secure environment, where the model reads the request and generates a response. The response is encrypted before being sent back to the browser. Under the described architecture, none of the intermediaries can read the content of the message.
Confidentiality has certain limitations. The template’s documentation notes that the application server can see the visitor’s IP address, while Nito can see the account, the model, the request time and the size of the encrypted message.
Developutors still rely on the code of the web page they provide, the hardware manufacturer’s attestation, and the published software running in that environment. Verification identifies the published release of the software; it does not establish that the release is free of errors.
Running the example requires a funded Nito account and an API key. Authentication, request rate limiting and bot protection remain up to the developers.
Developers can build agents based on verified results
In its discussion on X, the Foundation listed personal assistants, agents working with confidential business data, and private trading strategies as potential applications. It noted that developers can check the attestation status of the signing key before accepting signed results.
The launch announcement also lists other possibilities, including AI decision-making for prediction markets or insurance claims, accepting trade orders only from verified machines, and AI management decisions.
Solana’s privacy products span payments and token distribution
On November 10, Helius launched Solana Privacy Rings, offering encrypted balances and confidential transfers on mainnet. The initial version of this permissionless network hides assets and amounts while keeping sender and recipient addresses visible.
Earlier, on May 26, Umbra and Streamflow introduced a confidential token distribution system. The integration encrypts distribution amounts, recipient addresses and unlocking schedules, using Arcium’s encrypted computing infrastructure.
In the United States, Securities and Exchange Commission (SEC) Chairman Paul Atkins discussed financial privacy in a December 2025 speech. Speaking on his own behalf, he stated that zero-knowledge proofs and selective disclosure could help users demonstrate compliance without revealing their entire financial history.
