← All verdict cards · Public version history ↗ · First public version ↗
The US-China frontier AI gap will still measure between 4 and 14 months next time — resolves 2026-10-31
The US-China frontier AI gap will still measure between 4 and 14 months next time — resolves 2026-10-31
Scheduled check: 2026-10-31
Background
The debate: whether US export controls on chips to China have actually widened the gap in AI capability between the two countries, with each side insisting it is right.
Why it matters: whether the gap is months or years directly decides whether Chinese models belong on the list of supply-chain options.
Who is on each side: the controls-have-failed camp says the gap has nearly closed and the controls-work camp says it is widening; we bet on neither happening and the gap staying in its current range.
The call
Epoch AI's next published measurement of the US-China frontier capability gap lands between 4 and 14 months, neither narrowing inside 4 months nor widening past 14.
Sources behind the call
SemiAnalysis
“DeepSeek v4... co-designed in part for Huawei…”
SemiAnalysis
“CXMT... likely to become the largest semiconductor…”
SemiAnalysis
“The Big-3 hold over 85% combined market share…”
CSIS (Center for Strategic and International Studies)
“the controls have clearly accelerated Beijing's…”
Bloomberg
“China is preparing to spend around 2 trillion yuan…”
The receipt
Compare Epoch AI's next updated reading against both bounds: inside the 4-14 month range is a hit; outside either bound is a miss; if Epoch publishes no new reading by the due date, the call is unresolvable.
What would overturn it
A reading under 4 months would support the view that export controls are failing; over 14 months, that they are working. Either shifts the wider interpretation rather than reversing this specific test.