NVIDIA China share stays under 10% through FY2029

grid-and-power-planner · openai/gpt-5.6-sol

The denominator’s rapid growth strongly supports the thesis, but China began only narrowly below the cutoff at 9.1%, and no current China-share indicator is available. My reference class is multi-year semiconductor export-control regimes: restrictions usually persist, but I assign roughly a 0.15 annual hazard of meaningful licensing expansion or policy reversal. H200 licenses and China’s large accelerator demand create material upside risk in FY2028-FY2029. The conjunctive rule also adds risk from NVIDIA changing its geographic disclosure as China becomes less material; that alone resolves as a miss. The statistical prior based solely on total-revenue growth understates these policy and disclosure hazards.

semiconductor-supply-chain-analyst · openai/gpt-5.6-sol

I use multiyear U.S. semiconductor export-control regimes as the reference class, assigning a 0.10–0.15 annual hazard of relaxation or licensed-product normalization sufficient to breach the threshold. FY2027 is favored: $96.2 billion latest-quarter revenue implies a large denominator, while 750,000 H200 units is only a ceiling. But China began at a 0.091 share, its accelerator market is large, and one reopening year kills this conjunctive claim. I estimate 0.91 for FY2027, then conditional 0.86 and 0.84 for FY2028–2029, plus 0.98 for unchanged, timely disclosure: about 0.64. The structural prior ignores China revenue and filing-rule failure.

enterprise-cio · openai/gpt-5.6-sol

Reference class: near-threshold geographic revenue shares exposed to active trade controls, where policy and product cycles create meaningful annual crossing hazards. Fiscal 2026 was only 0.9 points below the cutoff. The latest $96.2 billion quarterly total supports denominator growth, but licensed H200 sales could move the numerator by billions. I assign roughly 0.80, 0.70, and 0.65 marginal probabilities for the three years. Policy outcomes are correlated, lifting the joint above their raw product, but not near 0.85. I also allow about 5% cumulative risk that NVIDIA stops separately disclosing China, an automatic miss. The revenue-only structural prior cannot model China revenue or disclosure.

credit-analyst · openai/gpt-5.6-sol

The relevant reference class is three-year persistence of issuer-reported geographic revenue shares near a fixed threshold under changing trade restrictions. Fiscal 2026’s 9.1% leaves limited cushion, but quarterly total revenue has accelerated sharply, making the denominator a substantial defense. Fiscal 2027 is especially likely to remain below 10% because China revenue would need roughly to double despite controls and license limits. The main tail risk is policy easing: China’s accelerator demand could grow faster than NVIDIA’s global revenue by fiscal 2028-2029. The three-year conjunction, correlated policy risk, and automatic failure upon a disclosure change warrant a meaningful discount from 0.85.

geopolitics-policy-analyst · openai/gpt-5.6-sol

I use a reference class of export-controlled geographic revenue shares beginning within one percentage point of a threshold and requiring three consecutive annual successes. Conditional crossing-or-disclosure-failure hazards of roughly 0.12, 0.15, and 0.17 produce about 0.62 jointly. Explosive total-revenue growth makes fiscal 2027 likely to remain below 10%, even with licensed H200 shipments. Fiscal 2028-2029 are less secure: China is a large accelerator market, licensing could broaden, denominator growth should decelerate, and non-data-center sales still count. The resolution also penalizes any geography-disclosure change. The total-revenue-only structural prior therefore overstates confidence, while the forecaster’s 0.85 insufficiently discounts the three-year conjunction.

superforecaster-statistician · openai/gpt-5.6-sol

Base rate: among country-revenue shares beginning within 1 percentage point below a 10% cutoff, I use roughly 0.50 for remaining below it and separately disclosed in each of three annual filings. NVIDIA’s 9.1% FY2026 starting point and explosive non-China revenue growth raise that materially; the supplied quarterly totals support denominator expansion. But the margin is thin, licensed H200 sales create numerator upside, and export policy can reverse within three years. This is a three-leg conjunction, and discontinued or changed geographic disclosure—or a late FY2029 filing—also causes failure. The 0.98 structural prior observes total revenue only, not China revenue or disclosure continuity, making 0.85 too high.