Nvidia Just Did Something It's Never Done Before. Here's Why It Makes the Stock a Buy.

Nvidia reported fiscal Q2 revenue growth of 106% to $96.2B, with adjusted EPS up 120% to $2.22 versus $2.09 and $92.3B sales consensus. Management guided fiscal Q3 revenue to ~ $108B (89% growth) and issued a rare forecast more than a year out, projecting 70% growth in fiscal 2028 (higher if not for supply constraints). Cash generation was strong with operating cash flow of $24.1B and free cash flow of $21.4B, while forward valuation was described as attractive at ~17x P/E on fiscal 2028 estimates—supporting a bullish stance after the earnings print.

Analysis

NVDA is increasingly trading less like a single-chip vendor and more like the toll collector on AI cluster expansion. The key implication is not just higher near-term earnings power, but lower perceived volatility of the demand curve; that can support a structurally higher multiple if delivery capacity keeps converting into backlog. In the ecosystem, the clearest secondary beneficiaries are networking/optics and advanced packaging names, while AMD and INTC remain pressured to prove that alternative silicon can displace an entrenched platform rather than merely participate in a rising tide.

The near-term risk is that the market extrapolates a supply-constrained guide too cleanly: when customers are capacity constrained, revenue can stay strong even as true end-demand normalizes, pushing disappointment risk out rather than removing it. Watch the next 1-2 quarters for hyperscaler capex cadence, AI cloud financing spreads, and any gross-margin compression from mix or pricing; those are the first places the story can break. Over 6-18 months, export controls, power availability, and customer concentration are the main variables that could turn a visibility premium into a durability discount.

Consensus is likely underweighting how much of the valuation is now justified by operating leverage plus scarcity, but overestimating how broad-based AI demand automatically maps to perpetual share gains. If model training gives way to more price-sensitive inference, pricing power can migrate from accelerators to networking, software, and infrastructure integrators. That argues for staying long the leader, but hedged against a sector-wide capex digestion because the whole AI stack still trades on one common denominator: continued spend growth.

Key Decisions for Investors

• Buy NVDA on 3-5% post-print pullbacks; 3-6 month horizon. The risk/reward is attractive if the market keeps paying for visibility rather than peak growth. Falsify on a material cut to forward revenue/gross-margin trajectory or a sharp deceleration in data-center growth next quarter.
• Pair trade: long NVDA / short AMD for the next 1-3 months. Thesis: NVDA remains the default platform for frontier AI spend, while AMD still needs multiple quarters of proof to win enough share to matter. Cover if AMD secures repeated large-design wins or NVDA loses pricing discipline.
• Long ANET / short SMCI over 3-6 months. As AI clusters scale, the networking layer should capture more wallet share than rack assembly, while SMCI carries higher working-capital and margin risk if customers slow deployments. Stop if server demand broadens faster than networking attach.
• If chasing strength, prefer call spreads over outright common only on a volatility pullback; avoid paying peak implied vol immediately after a euphoric repricing.