China's centralized aggregation grants massive yield in physical-world ground telemetry (IoT, Mobility, Clinical), while the US leads in frontier synthetic generation and open-web reasoning.
| Domain Sector | US Yield (T) | CN Yield (T) | Sovereign Lead |
|---|
The model simulates national AI training token yields across five distinct physical and computational domains. State pooling and privacy sliders rescale raw volume, structuring physical telemetry into effective training tokens. Selecting US Synthetic Leap boosts generated tokens and postpones the projected US training data wall from 2027 out to 2030.