
The Three-Way Race: Platforms, Frontier, and Infrastructure
The Chinese AI economy isn’t a two-player game. Three archetypes compete with different strategies and structural advantages.
Archetype 1: Big Tech Platforms
Players: Alibaba, ByteDance, Tencent, Baidu. Strategy: Distribution-first, zero-margin inference. Combined reach: 1B+ DAU. Stack coverage: L3-L7. Win condition: AI improves existing behavior.
Archetype 2: Frontier Research
Player: DeepSeek (hedge fund-backed). Strategy: Capability-first, efficiency breakthroughs. Intelligence Index: 68 (leading). R1 training cost: ~$6M. Win condition: AI creates new behavior.
Archetype 3: Infrastructure Sovereigns
Players: Huawei, Cambricon, Moore Threads, Biren. Strategy: Sovereignty-first, vertical integr — as explored in how AI is restructuring the traditional value chain — ation. Huawei Ascend: 50% share by 2026. Stack coverage: L1-L4. Win condition: Control supply chain.
The Barbelled Distribution of Value
Value concentrates at two poles — platform distribution and frontier research — while the middle gets squeezed.
Platform Pole: Monetizes via ecosystem economics. Can subsidize indefinitely. Owns consumer touchpoints.
Frontier Pole: Creates discontinuities via efficiency breakthroughs. Sets pricing pressure. Forces platform integration.
Squeezed Middle: Moonshot, Zhipu AI, MiniMax — neither distribution nor frontier. Becoming infrastructure — as explored in the economics of AI compute infrastructure — suppliers to Big Tech. The options: accept supplier role, pursue vertical specialization, or achieve frontier breakthrough.
If you are not at a pole, you are the product.
Scenario Probabilities by 2028
- 55% — Platform Consolidation: Big Tech absorbs disruption shocks. Distribution moats hold.
- 30% — Frontier Disruption: Agentic transition accelerates. Capability > distribution.
- 15% — Infrastructure Capture: Huawei vertical dominates. Chips → cloud → models.
This is part of a comprehensive analysis. Read the full analysis on The Business Engineer.









