LiveCambrian-leakage regime — Chinese variantWestern cycleAnchor-complete

Through 2040, the Chinese AI ecosystem continues to produce an accelerating diversity of models, deployed products, and new research directions?

As of 7 Oct · 00:37 UTCResolves 31 Dec 20404 movers0 pressure moments
Timeline A · State-directed bloom succeedsReality
55%
Crowd 55%Gap 0

By 2040, China's AI labs keep releasing new model families every few months, not just from Baidu and Alibaba but from dozens of smaller firms. Startups build specialized tools for manufacturing, finance, and medicine using Chinese-trained models. Research papers from Chinese authors top the major conferences, introducing techniques that competitors globally adopt and build on.

Timeline B · State-directed bloom plateausReality
45%
Crowd 45%Gap 0

By 2040, Chinese AI development plateaus into a smaller set of state-backed players. Model diversity shrinks as funding dries up and regulatory pressure focuses research on a few approved architectures. Startup formation slows; most new work recycles existing techniques rather than exploring novel directions.

crowd 55
The quiet paththe one nobody’s watching

A complete U.S. semiconductor equipment export ban to China would collapse both timelines' hardware foundation.

6%
The read
A different-shaped AI bloom in China — can it reach parity with the Western pattern?

China's state-directed AI bloom runs on different mechanics than the Western emergent pattern. Whether it reaches parity on applied + deployed dimensions by 2040, or plateaus before then, is the live question.

Where reality is heading

A, 80% bandCrowd today
Projected reality support to 31 Dec 2028A rises from 41% to about 52% with an 80% band of 33 to 71. B falls from 53% to about 44%. Quiet path near 4 to 6%. Crowd today: A 55%.100%7550250Crowd today · A 55%A ~52%B ~44%P ~4%NowJan 2027Jan 202831 Dec 2028
Reality crosses toward A in early 2028. Today’s crowd number sits at the top of A’s band, possible but optimistic. View as table

The movers, ranked by leverage

01
Research-direction outputNovel-direction publication + benchmark rate
leverage 0
Conditional
02
Startup formation + fundingNew-entrant pattern vs consolidation
leverage 0
Conditional
03
Compute-access distributionFrontier concentration vs distribution
leverage 0
Conditional
04
Regulatory environmentAcceleration vs dampening
leverage 0
Conditional

What would change our mind

Signals toward A
  • Novel-direction publication + benchmark rate continues to grow year-over-year
  • Startup formation + funding sustained above 2024 baseline
  • Compute access distributes to new entrants (via open-source models, cloud-platform competition)
  • Regulatory environment stays balanced (neither suffocates nor abandons the ecosystem)
Signals toward B
  • Novel-direction publication + benchmark rate visibly plateaus
  • Startup formation + funding consolidates to 2-3 major winners
  • Compute access concentrates further
  • Regulatory environment suffocates (or politicizes) the ecosystem

Evidence