Will leadership depend primarily on chips, electricity, algorithms, manufacturing, talent, capital, data, or commercial diffusion?
Buchanan's triad (data, compute, algorithms) frames the inputs; Ding's diffusion thesis counters that adoption capacity, not invention, decides hegemonic transitions — a lens under which the race looks very different.
View on the map → · Open in Browse →
The mid-year Chinese-model panic met its measurements: Kimi K3 is very good and below the frontier, and Epoch's index puts the Chinese lag at four to fourteen months. The sharper analyses argue the two countries are running different races — frontier capability versus deployment and diffusion — with China's open-weight strategy creating feedback loops that compute-focused controls don't reach, and electricity the input where China structurally leads.
While Kimi K3 is certainly a very good model, by the company's own admission, it is not at the frontier.
Sober read of the mid-2026 Chinese-model-quality panic: K3 is below the frontier, open-weighting it is rational for a lagging country, and China's incentives will converge with America's once its models get genuinely dangerous.
Since 2023, every model at the frontier of AI capabilities, as measured by the Epoch Capabilities Index, has been developed in the United States.
The measurement people cite for the US-China model-quality gap: a 4–14 month Chinese lag on the Epoch Capabilities Index, roughly mirroring the open-vs-closed gap.
If the models that matter most for industrial AI are small, specialized, and open, the current U.S. policy framework could be targeting the wrong layer of the competition.
A US government commission working paper identifying two compounding feedback loops — open-weight ecosystem diffusion and industrial deployment data — that give China advantages largely outside the reach of compute-focused export controls. Also central to the open-weights fight on 4.5.1.
When the commodity is cheap and abundant, strategic advantage shifts from who builds the best model to who deploys AI most effectively.
Diffusion-first argument applied to 2026: the US races for frontier capability while China competes on deployment across industry and the Global South — under which lens export-control denial looks counterproductive.
A Chinese Mythos-level model is inevitable and only months off.
Analysis of how Beijing would manage a frontier-level Chinese model, arguing the CAC's pre-deployment testing regime gives China a more orderly playbook than America's improvised response.
This growth in energy demand for AI is unlikely to be a constraint for China, given the country's historically rapid pace of overall energy expansion.
The 'electron gap' argument: electricity is the input where China structurally leads even while trailing on chips — the reference piece on energy as a decisive competition variable.