NVIDIA Corporation

Quarterly performance: strong growth + record demand (good) - Revenue $57B, +62% YoY, with record sequential growth of +$10B (+22%). - Q3 data center revenue: $51B, +66% YoY; compute +56% YoY (driven mainly by GB 300 ramp). - Networking: $8.2B, +162% YoY (NVLink scale-up + strong Spectrum X Ethernet / Quantum X InfiniBand growth). - Margins: GAAP gross margin 73.4%, non-GAAP 73.6% (above outlook); higher sequential margin driven by mix/cycle time/cost structure.

Visibility & forecasts: Blackwell/Rubin demand is high; $500B target reiterated with possible upside (surprising/positive) - Company states visibility to ~$500B in Blackwell + Rubin revenue from start of this year through end of calendar 2026. - CFO/CEO indicate they are on track and expect the number to grow, with ongoing order intake beyond the baseline (example cited: KSA announcement adding ~400k–600k GPUs over 3 years; Anthropic also referenced as incremental). - Q4 outlook: revenue $65B ± $2B (midpoint implies ~+14% sequential growth), driven by continued Blackwell momentum. - Gross margins in Q4: GAAP/non-GAAP expected ~74.8% ± 50 bps. - China constraint: like last quarter, no data center compute revenue assumed from China in guidance.

Supply vs demand & timing: GPUs are “sold out”/fully utilized; supply catch-up question remains key (risk/concern) - Management said clouds are sold out and the combined GPU installed base is fully utilized (including Blackwell, Hopper, Ampere). - CEO emphasized planning across the supply chain (TSMC packaging, memory, ODMs) but did not give a firm timeline that supply will fully catch demand. - Investor question focused on whether supply can catch up in 12–18 months; management response leaned on multi-layer AI demand drivers (accelerated computing + generative AI + agentic/physical AI all running on NVIDIA platforms) rather than explicit supply-fairness commitments. - Key implication: execution remains about whether production ramp can keep pace without extended shortages or lost revenue.