Who Will Make Money in the Age of Agents?
In the Agents era of blockchain, traditional value capture theories face challenges. The "Fat Protocol" theory, dominant since 2016, suggested protocols capture most value as their tokens are essential for network use. However, the proliferation of interchangeable L1s, L2s, and modular layers has eroded protocol scarcity and pricing power.
Conversely, the "Fat App" theory posits that applications capturing user relationships (like wallets and exchanges) become the primary value layer by controlling distribution and transaction flows. This aligns with the current "Great Repricing" cycle.
Agents disrupt this logic. As software users, they lack brand loyalty, prioritize cost and efficiency, and switch between platforms seamlessly. This undermines the front-end UX moats that "Fat Apps" rely on. The article explores several potential futures:
1. **Headless Applications:** Current leading apps could strip their front-ends and become backend API infrastructure for Agents, preserving their role.
2. **Protocol Resurgence:** If integration becomes trivial, Agents might bypass aggregators and interact directly with protocols, reviving "Fat Protocol" dynamics.
3. **Pricing Power Collapse:** Agents' rational, frictionless routing could commoditize the entire stack, compressing margins toward cost and leaving little profit for intermediaries.
4. **Unprecedented Activity:** Agents may enable new, high-frequency, machine-to-machine economic activities, expanding the total value pie even if margins are thin.
5. **A New, Unnamed Model:** Historically, major tech shifts (like the internet's attention economy) create unforeseen business models. The Agents era may spawn entirely new ways to capture value.
The most likely outcome is a coexistence where "Fat Apps" continue to serve human users valuing UX, while a separate, Agent-driven economy emerges governed by different rules—where loyalty is based on factors like liquidity, latency, and settlement guarantees rather than brand.
marsbit4h ago
