Guide
How to Evaluate AI Crypto Projects
Use a practical ten-point framework to assess AI crypto utility, token economics, usage, decentralization, security, competition, and market risk.
Evaluate the product, network, and token separately. A convincing AI demo does not prove that a decentralized network is necessary—or that its token captures value.
1. Define the actual product
Write down what a paying user receives in one sentence. Is it compute, inference, market data, an autonomous agent, model access, verification, storage, or a developer platform? If the answer stays vague after reading the documentation, the investment case is not yet clear.
2. Test whether blockchain is necessary
Look for a concrete coordination problem: independent suppliers, permissionless settlement, portable identity, transparent ownership, or verifiable rewards. “Decentralized” should identify which layer is decentralized and why users benefit.
3. Identify who pays
Separate subsidized activity from organic demand. Determine who purchases the service, what they pay, whether usage repeats, and whether customers would remain if token rewards declined.
4. Trace token value capture
Map the path from product usage to token demand. Does the token pay fees, secure the network, provide collateral, grant scarce access, or receive protocol value? Governance alone rarely explains durable demand.
5. Measure supply pressure
Review circulating supply, fully diluted valuation, unlock schedules, emissions, treasury holdings, and investor allocations. A useful product can still have poor market performance when new supply grows faster than demand.
6. Verify usage with more than one metric
Combine customers, revenue, paid jobs, active operators, retention, and fee volume. Avoid relying on raw transactions or social followers. For compute networks, ask how much capacity is actually used; for agents, ask which actions are independently executed on-chain.
7. Examine technical dependencies
List the centralized services required for the product to function: model APIs, cloud hosting, data providers, privileged keys, bridges, or a small validator set. A project can still be useful, but the dependency map should match its claims.
8. Review security and permissions
Agents and automated strategies can control wallets, approvals, and transaction execution. Check spending limits, human confirmation, key custody, contract audits, upgrade permissions, and the response plan for harmful or incorrect output.
9. Compare the competition
Compare the project with centralized AI providers, ordinary cloud services, and other crypto protocols. Lower prices funded by emissions may not remain lower. Network effects, proprietary data, distribution, and switching costs matter more than labels.
10. Separate fundamentals from narrative momentum
Use the AI Crypto Market dashboard to see whether price strength is broad or concentrated. Then compare it with the project-level evidence above. Momentum can be useful market context, but it does not validate product claims.
Red flags
- The documentation describes AI capabilities but never explains who pays.
- The token is unnecessary for the product’s main workflow.
- Usage depends primarily on emissions, quests, or circular incentives.
- “Autonomous agent” means a chatbot posting generated content.
- Market-cap comparisons ignore a much larger fully diluted valuation.
- Claims cannot be checked through code, transactions, customer evidence, or reproducible benchmarks.
Research on AI-themed crypto markets has also found substantial gaps between token narratives and demonstrated decentralization or economic utility. Skepticism is a useful part of the process, especially when financial claims exceed observable evidence.