The Next Uber Will Be an AI Agent, Not an App
Andreessen Horowitz’s crypto arm (a16z Crypto) is betting that the next wave of category-defining consumer platforms will be built around AI agents rather than human-facing apps. In a recent analysis, the firm argues that AI agents can borrow two powerful crypto-native primitives—intents and order-flow auctions—to reorganize two-sided marketplaces for complex transactions such as group dining, business travel, and procurement.
Rather than building another interface for humans to browse and book, a16z Crypto envisions agents that represent users’ preferences and constraints, then compete to fulfill those intents by bidding for order flow. This mirrors how decentralized exchanges aggregate liquidity and how searchers compete to fill user orders on-chain.
Why Crypto Primitives Matter for AI Agents
The core insight is that AI agents need credible neutrality, programmable incentives, and verifiable reputation to coordinate at scale. Crypto provides all three:
- Intents: Users express desired outcomes (e.g., “book a dinner for six under $400 total within 20 minutes of my office”) rather than navigating menus. Agents compete to satisfy the intent.
- Order-flow auctions: Agents bid for the right to fulfill intents, creating a market for execution quality. This aligns incentives between users, agents, and underlying service providers.
- Fulfillment reputation: On-chain attestations and staking mechanisms can build trust in agents’ reliability, creating a long-term moat.
a16z Crypto recommends starting with a narrow vertical—single-city restaurant dining—where transaction complexity is manageable but still involves multiple variables like availability, pricing, dietary restrictions, and group coordination. From there, the model can expand to travel and procurement.
Monetization and the Path to a Moat
The proposed business model centers on paid order flow: service providers pay agents for qualified demand, and agents pass savings or convenience to users. Over time, the winning platform becomes a trust layer—a network where agents’ performance history, dispute resolution, and reputation are transparent and portable. That trust infrastructure, rather than the interface, becomes the durable competitive advantage.
This vision challenges the conventional app-centric playbook. If agents become the primary interface for commerce, the value accrues to the coordination layer—the marketplace protocol and reputation system—not the front-end. That is precisely where crypto’s open, composable infrastructure shines.
Implications for Builders and Investors
For crypto builders, the message is clear: the intersection of AI and crypto is not just about decentralized compute or model tokenization. It is about using blockchain rails to coordinate economic activity among autonomous agents. Projects that provide identity, reputation, payment, and dispute resolution for agents could become foundational.
For investors, the a16z Crypto thesis suggests watching for startups that combine agentic AI with marketplace mechanics in high-friction verticals. The next Uber may not be an app at all—it may be a network of agents quietly optimizing the economy in the background.
Forward Look
The transition from human-driven apps to agent-driven networks will not happen overnight. Regulatory clarity around agent liability, data privacy, and financial transactions will be critical. But if a16z Crypto is right, the companies that build the trust and coordination infrastructure for machine-to-machine commerce will define the next decade of consumer technology—and crypto will be the connective tissue.

