Nvidia CEO Signals Strong Momentum for Shareholders

Nvidia reported fiscal Q2 2027 adjusted EPS of $2.22 on $96.22 billion in revenue, above consensus estimates of $2.10 and $92.17 billion, respectively; revenue grew 106% year over year. CEO Jensen Huang indicated roughly 70% sales growth for fiscal 2028 versus the 44% analyst consensus, while projecting 72%-73% gross margins despite supply constraints. Nvidia also expects its Vera Rubin GPUs to be its fastest-ever product ramp, with revenue potential of $40 billion per gigawatt of AI compute versus $25 billion for Blackwell, and forecasts CPU revenue will more than double next fiscal year.

Analysis

The investable implication is not simply higher GPU volumes: NVIDIA is expanding dollar content per AI data-center build through CPUs, networking and systems, which raises switching costs and shifts the debate from accelerator share to share of total AI infrastructure spend. That creates the clearest competitive pressure on AMD in accelerators and CPUs, Intel in server CPUs, and potentially Arista/Marvell/Broadcom where NVIDIA bundles networking more tightly with compute. The key incremental earnings variable over the next 1-3 quarters is whether non-GPU attach revenue grows faster than compute revenue, sustaining mix-driven margin resilience even as new-platform ramps normally dilute gross margin.

The guidance optimism should be discounted until hyperscaler capex budgets, backlog conversion and customer prepayments corroborate it; management commentary is not equivalent to a contracted demand schedule. Supply constraints can preserve pricing near term, but they also defer revenue and create a vulnerable setup if any major cloud customer pauses deployment after securing capacity. A meaningful downward revision to 2027 cloud capex plans, inventory build at ODMs, or gross margin failing to recover after the transition would challenge the premium multiple within days; a 6-18 month risk is that AI inference economics favor lower-cost ASICs or internally designed silicon over NVIDIA's full-stack offering.

Consensus may be underestimating NVIDIA's ability to monetize the CPU/networking layer, but is likely over-extrapolating each generation's revenue-per-gigawatt economics without adjusting for customers' fixed data-center power and capital budgets. Higher NVIDIA content per deployment can crowd out spend on adjacent components rather than expand total spend, making the cleanest relative expression long NVDA against exposed merchant-silicon competitors rather than an outright sector beta trade. NFLX has no direct read-through beyond broad AI-investment sentiment and should not be treated as a beneficiary.

Key Decisions for Investors

• Maintain or add a 1-3 month long NVDA / short AMD pair: NVIDIA's platform attach opportunity can drive upward estimate revisions while AMD faces both GPU and server-CPU share pressure. Target 10-15% relative outperformance; exit if NVIDIA's next reported data-center growth decelerates materially without a corresponding gross-margin recovery, or AMD demonstrates sustained MI-series supply and customer wins.
• Watch for a long NVDA / short INTC relative trade into the next server-demand update, contingent on evidence that NVIDIA CPU revenue is converting into shipments rather than design-win commentary. The thesis is strongest if Intel's Data Center and AI guidance weakens; avoid initiation if Intel shows sequential server-unit recovery and gross-margin stabilization.
• Reduce broad AI-hardware basket exposure rather than chase a post-results move. Use the next hyperscaler earnings cycle and supplier commentary as the catalyst window; any cloud-capex revision lower or evidence of ODM inventory accumulation would likely hit high-multiple AI infrastructure names together, regardless of NVIDIA's relative execution.
• Set monitoring triggers: hyperscaler capex guidance, NVIDIA networking/CPU revenue disclosure, and post-ramp gross-margin trajectory. If margins remain below management's expected recovery range after the new-platform ramp, trim NVDA exposure because the market will reassess whether systems mix is structurally less profitable than assumed.