At first glance, the token economy looks like the beginning of a new era. I think we are already at the beginning of its end.
The token economy is a transitional phase, not a destination. To understand why, we need to begin somewhere seemingly unrelated: geopolitical chokepoints.
The new chokepoints
Every civilization organizes itself around geographic bottlenecks: the chokepoints. The British Empire understood maritime chokepoints better than anyone. The United States inherited much of that strategic architecture postwar.
A chokepoint is a narrow passage: straits, canals, sometimes mountain passes. Malacca, the Suez Canal, the Panama Canal, the Strait of Hormuz and dozens of lesser-known corridors still define the map of global trade.
The map of chokepoints largely defines today's global trade, macro and geopolitical orders. But that map is becoming obsolete.
Different maps being drawn
Chancay, a new port in Peru, can send Brazilian soy and Chilean copper straight to Shanghai, bypassing Panama and Los Angeles. The Arctic Route can cut the Suez load. Rail routes across Kazakhstan and Russia, the Iron Silk Road and the Middle Corridor create new paths through the interior.
Industries are beginning to build their own sector-specific routes, from electric vehicles and lithium to semiconductors. Material mined in Kazakhstan can move to processing hubs in China before finished EVs travel directly to the European market.
Beyond borders, another map of chokepoints appears: rare earth processing, semiconductor fabrication, advanced packaging and ultra-high-voltage transmission technology.
The inference economy
If energy is the new chokepoint, the obvious move would be to double down on the energy-draining token economy. Instead, China appears to be making a different bet: inference chips, precision compute allocation and efficiency over brute force.
China's AI strategy is developing its own doctrine language. The counter to a training-parity strategy is not simply matching GPU for GPU. It is building inference chips, running models cheaply at scale and tailoring silicon to software.
The transformer is done evolving. The model is the chip now.
Efficiency over hyperscale
Hardware is only half the battle. Sparse expert routing, hardware-native quantization and end-to-end reinforcement learning point toward a different theory of progress: do less computation, but make each unit of computation more useful.
Efficiency is the new hyperscale.
Data over compute
For two years, the industry measured AI progress in cluster sizes. The next contest may be measured in the quality of the environments, problems and feedback used to train intelligent systems.
The center of gravity is moving from raw compute toward the relationship between silicon, training method, input data and the energy system beneath them.