AI Agents Need Data Before They Can Transact
AI agents are getting wallets, stablecoin balances, and the ability to sign transactions onchain. The infrastructure for autonomous economic actors is being built in real time: Coinbase shipped Agentic Wallets and the x402 payment protocol, MoonPay launched non-custodial agent infrastructure, and protocols like deBridge now support AI-initiated cross-chain execution. Agents can hold capital, swap tokens, and settle payments without human intervention.
The momentum is not confined to crypto-native companies. In its 2025 annual letter, Stripe reported that stablecoin payment volume doubled to roughly $400 billion, with 60% estimated to represent B2B transactions. Stripe launched machine payments for developers to charge agents directly using stablecoin micropayments, and partnered with OpenAI to release the Agentic Commerce Protocol, a standard for AI agents to transact on behalf of buyers. Co-founder John Collison put it plainly: a "torrent" of autonomous AI agents is coming, and they will handle commerce using stablecoins on high-throughput blockchains.
But there is a meaningful gap between "can transact" and "should transact." An agent with a wallet and no context is just an automated spender. The question the industry has not yet answered is more fundamental than custody or settlement: where do agents get the verified data they need to make good financial decisions?
The Wallet Problem Is Solved. The Data Problem Is Not.
Most of the current infrastructure buildout focuses on giving agents the ability to move money: wallets, signing, custody, settlement. These are necessary primitives, and they are arriving quickly. But what remains missing is the layer that sits between an agent's intent and its execution: the data that informs whether a transaction makes sense in the first place.
Consider what a sophisticated financial agent actually needs to do before executing a trade, rebalancing a portfolio, or entering a lending position:
→ Query the historical behavior of an asset or position over days, weeks, or months
→ Aggregate onchain activity across protocols and chains to assess risk
→ Verify that the data underlying a decision has not been tampered with
→ Access offchain inputs (reserve levels, cash flows, NAV calculations) alongside onchain state
→ Execute deterministic logic over those datasets and deliver the result to a smart contract
None of this is possible with current-state data alone, and none of it is trustworthy without cryptographic verification.
Smart contracts today can enforce rules based on current state, but they cannot access or compute over historical data. The same limitation applies to any agent interacting with those contracts. An agent evaluating a lending position cannot ask: how has this borrower's collateral ratio behaved over the last 30 days? An agent managing a tokenized portfolio cannot compute rolling NAV based on verified cash-flow data. An agent distributing rewards cannot prove that its eligibility calculations are accurate and auditable.
The data layer for agentic finance does not yet exist in most infrastructure stacks, and until it does, agents will remain constrained to the same narrow design space as the contracts they interact with.
Why Verified Data Is Non-Negotiable for Autonomous Agents
When a human trader makes a decision based on a dashboard, there is an implicit trust layer: the person evaluates the source, cross-references information, and applies judgment. Agents do not have that luxury. They execute based on inputs, and if those inputs are wrong, stale, or manipulated, the agent acts on bad data at machine speed.
Data verification a prerequisite for agentic finance. For agents to operate in financial markets with any degree of reliability, the data they consume must be provably correct and provably untampered. Without that guarantee, autonomous agents introduce systemic risk rather than efficiency.
The problem compounds in DeFi, where agents interact with composable protocols across chains. A single agent managing a cross-protocol strategy might need to read Ethereum lending data, check Base stablecoin flows, verify offchain reserve levels, and compute an aggregated risk score before executing a position. Each of those data sources needs to be verifiable, and the computation over them needs to produce a cryptographic proof that a smart contract can check. The same challenge applies to the agentic commerce layer that Stripe and others are building: when agents transact using stablecoins at scale, the financial logic informing those transactions needs to be grounded in data that both parties can trust.
As agents begin managing meaningful capital, the integrity of the data pipeline becomes less of a technical nicety and more of a financial infrastructure requirement.
Where Space and Time Fits
Space and Time, the data blockchain securing onchain finance, provides the verifiable data infrastructure that agents need to move from simple transactions to informed financial decisions.
Space and Time indexes onchain data from major blockchains, stores it across a decentralized network of database validators, and makes it queryable with SQL. The results of those queries are secured with cryptographic proofs, which means a smart contract (or an agent interacting with one) can verify that the data and the computation are correct before acting on them.
This creates a fundamentally different data pipeline for agentic finance:
Historical data becomes accessible. Agents can query position behavior, protocol activity, and asset performance over any time window, not just the current block. Space and Time stores the full history of indexed chains in tamperproof tables, making historical analysis a native capability rather than a custom build.
Computation becomes verifiable. When an agent computes an aggregation, a risk score, or an eligibility check, Space and Time generates a cryptographic proof of the result. The smart contract receiving that result does not need to trust the agent or the data source: it verifies the proof directly.
Offchain and onchain data converge. Many agentic workflows require joining blockchain data with offchain inputs: reserve levels, cash-flow data, compliance parameters, market feeds. Space and Time supports secure offchain data tables alongside indexed blockchain data, enabling agents to operate across both worlds with the same verification guarantees.
Data-driven logic executes deterministically. Rather than relying on agents to build and maintain their own data pipelines, Space and Time enables scheduled, repeatable query logic that produces consistent outputs. Deterministic rollups, aggregation logic, and rule-based computations run inside the network, delivering results that agents and smart contracts can consume directly.
The Agent Stack Needs a Data Layer
The infrastructure stack for agentic onchain finance is taking shape: wallets provide custody and signing, payment protocols handle settlement, and execution layers route transactions. But between intent and execution, there is a critical dependency that remains largely unaddressed: verified, queryable, historical data that agents can trust and smart contracts can confirm.
Without that layer, agents are confined to the same narrow design space as current smart contracts: single-chain, current-state, trust-dependent. With it, agents gain access to the full depth of onchain and offchain financial data, computed and proven in a way that supports sophisticated financial decision-making.
Space and Time is built for this requirement. As AI agents become economic actors in onchain markets, the data they consume will determine the quality of the decisions they make. Verifiable computation over historical and real-world data is the foundation of this economy.
