AIStablecoin
|5 min ReadA16Z: AI x crypto crossovers
Maya Chen
Senior Analyst
Published
Jan 16, 2026
The open web is already bending under the weight of AI. As more user journeys collapse into a single prompt bar, the business model that funded “free” content starts to wobble. The big question is not whether AI changes the internet, it is whether the result is an open network or a tighter maze of paywalls and platform control.
a16z crypto’s latest framing is simple: crypto is not a side quest, it is a set of architectural and economic tools that can push back against the centralizing gravity of AI. Their core claim is that blockchains can create credibly neutral rails for identity, ownership, and payments, so new AI-native services do not have to live inside a few closed stacks.
The prompt bar era creates a new control point
When discovery, navigation, and even transactions get mediated by AI interfaces, the “front door” to the internet shifts. That shift concentrates leverage. Whoever controls the assistant layer can decide what gets cited, what gets routed, what gets paid, and what gets blocked.
That is why the piece keeps returning to incentives. If AI systems can pull value from the open internet without compensating sources, the default equilibrium becomes closure: more paywalls, more bot blocking, more legal fights, and more creators deciding it is not worth publishing openly. Crypto enters the story as a way to program revenue sharing and user custody into the rails themselves, instead of relying on policy or platform benevolence.
Why blockchains show up in the AI stack
The argument is not “put everything onchain.” It is that some primitives are missing in an agentic world: portable context, a universal identity layer for agents, credible proof of personhood, and cheap, high-frequency payments. In a world of many agents talking to many services, closed APIs and one-off integrations do not scale. Open standards do, but only if there is a neutral coordination layer that no single company can unilaterally rewrite.
Blockchains are positioned as that layer because they can be permissionless, interoperable, and upgradeable, while still giving users custody and enabling composability across ecosystems.
Eleven crossovers that move from theory to buildable systems
The piece groups the crossovers into four themes: identity, decentralized infrastructure, new incentive models, and owning future AI. The goal is to make the overlap less hand-wavy by pointing to concrete system shapes being built today.
First, persistent data and context in AI interactions: the idea is to treat key context as a portable asset, so users do not have to rebuild preferences and project state every time they switch tools. Second, universal identity for agents: agents need a “passport” that works across email, Slack, marketplaces, and other agents, acting as wallet, capability registry, versioning, and reputation. Third, forwards-compatible proof of personhood: as bots flood feeds and deepfakes get cheaper, reusable proof that you are a real human becomes baseline infrastructure, ideally privacy-preserving and not controlled by a single issuer.
On infrastructure, DePIN for AI targets the physical bottlenecks, especially compute and energy, by aggregating idle chips into permissionless marketplaces and enabling distributed training, fine-tuning, and inference. The next layer is agent-to-agent interaction rails and guardrails: standardized workflows, payments between agents, and protocol-level constraints that keep agents aligned with user intent, instead of fragile point-to-point API wiring. Then there is the “vibe-coded apps” problem: AI-assisted development accelerates shipping, but increases entropy across dependencies and interfaces, pushing demand for synchrony layers that keep software compatible as it mutates in real time.
On incentives, the most direct wedge is micropayments and revenue sharing. If AI-driven actions lead to conversions, the sources that informed those decisions need a programmable cut. That requires tiny, frequent payments and attribution mechanisms, something crypto systems, especially low-fee rails, try to make feasible at scale. Adjacent to that is IP and provenance registries: public, programmable ownership records that make licensing and remixing legible to both humans and machines, with protocols aiming to turn AI’s remix dynamics into monetizable distribution instead of pure extraction.
Then comes the webcrawler reality. Bots are a massive share of traffic, and many ignore robots.txt, leaving sites paying the infrastructure bill for scraping. The proposed middle path is not just blocking, but paid access: crawlers negotiate and pay, humans prove personhood and access freely, so creators get compensated at collection time without shutting the door on real users. Advertising also gets reframed: “tailored, not creepy.” The promise is that AI agents can personalize based on user-defined preferences without globally exposing user data, and users can be compensated for engagement, but only if payments are cheap and privacy-preserving verification, such as ZK proofs, is practical.
Finally, owning the future of AI focuses on companions. As AI becomes a relationship, not just a tool, ownership and control of that relationship becomes a high-stakes question. The claim is that censorship-resistant hosting and user-controlled wallets, made easier through better UX patterns like embedded wallets, passkeys, and account abstraction, may be the most realistic path to user-owned companions for most people who cannot run models locally.
The throughline is consistent: AI shifts power toward whoever controls the interface and the incentives. Crypto’s best case is not ideology, it is plumbing, identity, payments, and ownership that remain portable when the internet gets recompiled into agents.
Disclaimer: This document is intended for informational and entertainment purposes only. The views expressed in this document are not, and should not be taken as, investment advice or recommendations. Recipients should do their own due diligence, taking into account their specific financial circumstances, investment objectives and risk tolerance, which are not considered here, before investing. This document is not an offer, or the solicitation of an offer, to buy or sell any of the assets mentioned.