Space and Time: Verifiable Onchain Data for AI Agents with Proof of SQL

Space and Time's SXT Chain and Proof of SQL technology are now live for over a year, giving AI agents and smart contracts a way to query Ethereum data with cryptographic proofs. For builders, this means you can gate agent actions based on verifiable onchain data, not just trust a centralized API. The system is already in use for capital allocation and vault management, though the planned DataFi marketplace has been paused.

SXT Chain: A Purpose-Specific Chain for Verifiable Queries

Space and Time did not build a general L1. Instead, it built SXT Chain as a purpose-specific chain of nodes that sit next to Ethereum, verify data, and return query results to smart contracts. Co-founder Scott Dykstra describes it as "almost like an oracle for querying data" in a recent interview. The architecture is designed to solve a specific problem: allowing smart contracts to query verifiable data without trusting a centralized oracle.

Proof of SQL: Sub-Second Cryptographic Proofs

Proof of SQL lets Prover nodes handle queries and generate proofs in under a second for most workloads. Common queries that run constantly, like checking collateral before a trade, stay sub-second and are cheap to execute. Complex aggregations over months of Ethereum history can take longer, but those are executed on demand, not continuously.

Live AI Agent Use Cases: Capital Allocation and Vault Management

AI agents are already using verifiable onchain data in production. Dykstra cites two live examples:

• Capital allocation to sub-accounts: A smart contract queries Space and Time to check an agent's trading performance over a time period, then rescinds or adds more capital to that agent's sub-account.
• Vault management: Before an agent moves assets into a yield vault, execution is allowed only if a proof shows the vault has sufficient liquidity, has stayed below a maximum historical drawdown, and has not exceeded exposure limits.

These are early but concrete. Dykstra notes that trading onchain with AI agents is still in its infancy, and more use cases will emerge.

Validator Economics and the DataFi Pause

Space and Time reduced its validator set from the initial 34 to 21 active validators, which increased the target staking reward to about 9.7% annualized. The planned DataFi model, where dataset contributors earn fees when their datasets are queried, has been paused. Dykstra explains that most customer datasets are too specific to their own use case to be valuable to others, and onchain data marketplaces have historically struggled to get adoption. If demand changes, Space and Time will revisit the model.

Dreamspace and the Limits of AI-Generated Code

Dreamspace, an AI-assisted code generation tool, offers a first-pass audit using prompts collected from top auditing teams. But Dykstra is clear: this is not a replacement for a professional audit, especially for sensitive contracts. Space and Time's core engineering focus remains on Proof of SQL and its five years of cryptographic research, not on Dreamspace.

The Bottom Line for AI Builders

For teams building AI agents that need to interact with onchain data, Space and Time provides a way to get cryptographically verified data without trusting a centralized intermediary. The technology is live, the use cases are real but nascent, and the company has been honest about what works and what doesn't. The DataFi pause is a reminder that not every data marketplace model succeeds, but the core verifiable onchain data capability is proven and available today.