TL;DR: As AI agents begin to think, act, and transact independently, the question becomes how AI can safely participate in the economy. Blockchain enters the picture, providing the coordination layer needed for autonomous agents to become trusted economic actors.
In this article, I break down a16z's AI × Crypto thesis: how Know Your Agent (KYA), and cryptographic trust let AI agents work together. This piece also discusses why micropayments are important for a sustainable AI economy and which projects and infrastructures are worth watching.
Thesis 1: Blockchain can serve as an infrastructure layer for AI models and agents to collaborate effectively.
AI is approaching the stage where it can solve problems that only a few specialists in the world can solve. Recently, ChatGPT 5.2 successfully solved a mathematical problem that only a few hundred individuals worldwide are capable of solving.
We used to blame AI for hallucinations.
As AI improves, these missteps can help it combine ideas and link things like humans who brainstorm. At scale, unlocking this creativity requires moving beyond a single model to layered systems where one AI freely generates ideas, second critiques them, third refines the best parts, and fourth validates the final result.
But, once multiple AIs are run together, two fundamental problems emerge:
Interoperability
Accountability.
Different models operate in different formats, which makes coordination difficult without a shared language or control layer. When one AI comes up with an idea, another AI improves it, and a third AI checks it, it's hard to figure out who gets credit, who gets paid, and who is responsible.
This is where crypto and blockchain naturally fit in, not as intelligence but as infrastructure by recording who did what, when it happened, and how much each contributor added. Through verifiable logs, hashes, attestations, and automated incentives, crypto can act as the accounting and coordination layer that allows diverse AI systems to collaborate.
Watchlist
1. @Covalent_HQ is creating a modular data architecture that allows AI agents to collaborate using shared and verifiable blockchain data. Multiple agents collaborate on complex tasks using their AI Agent SDK and Zero-Employee Enterprise workflows, while Block Specimens and the GoldRush API ensure chain and tooling interoperability. This positions blockchain as the foundation for data availability, verification, and incentives.
2. @AlloraNetwork is developing a decentralised coordination layer for collaborative AI, where multiple models work together on very specific tasks to get better results. Allora uses crypto to coordinate participation, verify contributions, and make sure that different AI agents work together in a way that makes the system smarter over time.
3. @questflow is building an on-chain orchestration layer for the multi-agent economy, where autonomous AI agents can talk to each other, coordinate actions, and complete entire workflows together. Rather than just having isolated agents that do one thing. Questflow's Multi-Agent Orchestration Protocol (MAOP) lets agent swarms work together to reason, decide, act, and settle payments.
4. @Gaianet_AI routes, load-balances, and serves requests for many independently run AI agents in shared domains. By standardising runtimes (WasmEdge), OpenAI-compatible APIs, and agent composition (LLMs+ RAG + tools), Gaia solves interoperability between heterogeneous agents at scale. The network has 700,000+ nodes and 29+ trillion inference throughputs, demonstrating real-world use. Rather than relying on provider trust, Gaia uses protocol-level mechanisms such as Onchain IDs, escrow contracts, and staking to introduce accountability into AI agent execution.
5. @SentientAGI is building the GRID, an open intelligence network where 100+ models, agents, data sources, tools, and compute providers work together as a single system. Instead of a monolithic AGI, GRID routes each query across the most relevant specialized intelligences and merges their outputs into a coherent result.
The network is live with 110+ partners and uses a token-based model where staking and real usage direct rewards to valuable artifacts, aligning funding with utility. By letting agents trade directly in SENT, crypto becomes the coordination and incentive layer that makes open, networked intelligence sustainable at scale.
Aside from the above project, I discovered two research papers that are really interesting. If you want to learn more about research and dive deep into these spaces, you should definitely check them out.
1. Intelligent System of Emergent Knowledge (ISEK) : ISEK proposes a coordination fabric where human and AI agents don’t just execute tasks, but discover each other, negotiate roles, form temporary teams, and settle outcomes through a native protocol loop (Publish → Discover → Recruit → Execute → Settle → Feedback). Trust, memory, and incentives are first-class primitives: Agents have verifiable identities (Agent Cards / NFTs), build multidimensional reputations, and trade value through tokenized micropayments based on how well they do their jobs.
2. LOKA Protocol: A Decentralized Framework for Trustworthy and Ethical AI Agent Ecosystems
Itl is an academic proposal for governing large-scale AI agent ecosystems. It introduces a layered architecture where agents have self-sovereign identities (DIDs + Verifiable Credentials), intent-aware communication, and a decentralized ethical consensus mechanism that lets agents reason about what they should do, not just what they can do. LOKA explores how accountabilityand ethics can be embedded directly at the protocol layer using on-chain logs, reputation-weighted consensus, and even post-quantum cryptography
Thesis 2: AI agents need identity, not more intelligence. Know Your Agent (KYA) is the missing primitive that makes them trustworthy economic actors.
Today, AI agents are already acting in the real economy. They make payments, book services, trade assets, negotiate deals, and operate critical financial infrastructure through APIs, bots, scripts, and automated systems. These agents are smart enough to work; intelligence is no longer a barrier. Identity and trust are the issues. When an agent pays, orders, or signs a contract, no one knows who it belongs to, what it can do, or who is responsible if something goes wrong. Therefore, websites and merchants block them by default with CAPTCHAs, IP bans, and bot protection.
The solution is Know Your Agent (KYA). Agents need cryptographic identity and verifiable credentials, just like humans need legal identity. Each agent must have signed keys proving its creator, what it represents (person, company, or DAO), its constraints, and its liability if it harms. These credentials explicitly state the agent's spending, trading, and data access limits, making responsibility clear.
WatchList
1. @billions_ntwk is building Know Your Agent (KYA), using the Agent JS SDK, agents generate their own DIDs, prove control via cryptographic signatures, and manage keys through a modular KMS, enabling identity, accountability, and reputation for agents. Already on board 2,372,153+ users
In partnership with Privado ID (formerly Polygon ID), Billions leverages zero-knowledge, self-sovereign identity to enable private verification across any service, device, or protocol.At the core lies $BILL, a fixed-supply ERC-20 utility token powering the trust economy, where Network Growth → Verification Activity → Revenue → Onchain Buybacks → Reduced Supply → Value Appreciation → Network Growth, aligning real usage with long-term value accrual.
2. @cheqd_io is building trust infrastructure for the agentic economy by turning Know Your Agent (KYA) into real, shippable infrastructure. Through Agentic Trust Solutions, AI agents get verifiable DIDs, fine-grained credentials, permissions, and accreditations, all anchored in tamper-proof Trust Registries.
With MCP (Model Context Protocol) servers, agents can read/write identities, issue and present verifiable credentials, and prove who built them, what they’re allowed to do, and why they’re trusted
3. @VouchedID is building a Know Your Agent (KYA) stack focused on security, accountability, and compliance for AI agents. Through MCP-I (Model Context Protocol – Identity), agents get cryptographically verifiable identity, delegated authority from humans, action limits based on context, and full audit trails.
This stack is reinforced by knowthat.ai, a public agent reputation registry, and the Vouched Agentic Bouncer that blocks unauthorized or impersonated agents, making autonomous AI safe to deploy in regulated, real-world environments.
4. ERC-8004 (Trustless Agents) is a proposed Ethereum standard (EIP) that is not yet a finished protocol. Its main goal is to make Know Your Agent (KYA) possible at the protocol level. It defines how AI agents can have verifiable on-chain identity, reputation, and execution proofs. This lets users and services figure out an agent's authorization and trustworthiness without having to use centralized platforms. The EIP is being actively designed and discussed by the Ethereum Foundation team, with contributors from Coinbase, MetaMask, and others.
Thesis 3: Blockchain can enable real-time, usage-based micropayments and nanopayments that automatically compensate creators when AI agents or tools use their content, ensuring fair and transparent revenue sharing.
AI tools like ChatGPT, Claude, and Copilot simplify user life but quietly disrupt the open web's revenue model. The web survives on ads, subscriptions, and paid views, yet AI has changed the value loop entirely:
Before AI: you searched → clicked websites → websites earned money.
Now: you ask AI → it reads websites → gives answers → websites lose traffic and income.
This creates an “invisible tax,” where AI consumes information without paying the creators who produced it. If this continues, websites will lose traffic, ad revenue will collapse, creators will stop publishing, and the open web will shrink, ironically leaving AI with less fresh high-quality data to learn from. Although laws may help, they move too slowly, making technical, incentive-aligned solutions urgent.
What needs to change is a shift to usage-based compensation, where every time an AI uses information, the creator is automatically paid in real time. Content would be paid per AI usage, like Spotify pay per stream and YouTube per view, instead of fixed licensing deals.
This works through micropayments and nanopayments, where AI credits multiple sources for an answer and splits tiny payments proportionally using math, not manual deals. For example: 20% from Site A, 30% from Site B, 50% from Site C and paid proportionally.
This is where blockchains and crypto come in; By integrating these automated payments directly into the web via blockchains and smart contracts, AI can continue to provide convenience while fairly compensating the creators on whom it relies.
Watchlist
1. @catena_labs is building an AI-native financial institution designed specifically to let AI agents participate directly in the economy. Through its open-source Agent Commerce Kit (ACK), Catena Labs provides AI agents with wallets, verifiable identities, payment rails, and rule-based spending controls, enabling them to make and receive payments autonomously. By supporting stablecoin payments, micropayments, and agent-to-agent transactions on blockchain testnets, ACK makes it possible for AI agents to automatically compensate other agents or human creators whenever their data, content, or services are used.
2. x402 embeds nano-payments directly into standard HTTP requests with near-zero friction to allow AI agents to pay for content, APIs, and compute immediately. @GoKiteAI transforms this payment primitive into a full execution layer, creating a blockchain where autonomous AI agents can reliably settle pay-per-use transactions at scale. Kite allows AI agents to automatically pay creators, services, and other agents when consumption occurs using x402-style flows, agent-native identity and stablecoin settlement.
3. @AIsaOneHQ is building an AI-native payments and billing layer that lets AI agents pay only when they act, using a single account, token, and API. It enables on-demand micropayments with per-request and per-token metering, powered by low-latency blockchain infrastructure and emerging agent-present payment standards.
With 10.5M+ x402 transactions already processed (≈16% of network activity, primarily on Base) and expansion planned to Solana and Polygon, this demonstrates that AI-native micropayments can operate reliably at scale.
