Nvidia's Valuation Just Hit Lows Not Seen Since 2019, Even as Its Profits Doubled

Nvidia’s fiscal Q2 results delivered record revenue of $96.2B (+106% YoY, +18% QoQ) and adjusted EPS of $2.22 (+120% YoY), beating consensus ($92.3B revenue, $2.09 EPS). Data center revenue jumped to $89B (+117% YoY) and management reiterated demand for next-gen processors while setting Q3 revenue guidance of $108B (+89% YoY) with gross margin dipping to ~74% (midpoint). Shares rose more than 4% after hours as investors also took comfort from an expanded AWS AI-infrastructure collaboration targeting deployment of more than 3M Nvidia GPUs.

Analysis

The real signal is not the beat; it is that hyperscaler demand is still being funded at a pace that forces the ecosystem to stay in build mode. That keeps NVDA in the role of a toll collector on the AI capex cycle, but the second-order winners may be the companies that turn GPU deployment into utilization and service revenue faster: AMZN’s AWS, plus the plumbing names around memory, networking, power, and cooling. The risk is that investors keep extrapolating unit demand while ignoring that the next leg of upside has to come from monetization, not just procurement.

Near term, the stock can stay strong for days on reflexive positioning and momentum, but the 1-3 month setup is all about whether this demand converts into durable guide raises or merely a larger backlog of future shipments. If gross margin stalls around the low-70s or the next guide implies growth is decelerating faster than expected, the market will start to question whether the AI spend curve is being pulled forward rather than expanded. That would hit the whole AI complex, not just NVDA, because the trade has become crowded around the same scarcity narrative.

The contrarian view is that the consensus may be underestimating concentration risk: a few hyperscalers can keep the headline numbers strong while still tightening procurement, changing mix, or substituting internal silicon over time. Six to eighteen months out, the bigger issue is multiple compression if earnings growth normalizes before the market broadens the customer base. In that scenario, the winners shift from the pure hardware beta names toward platforms that capture workflow and cloud utilization economics, especially AMZN.

Key Decisions for Investors

• Stay long NVDA on pullbacks, not momentum-chase after the gap; use a 3-5% retracement as preferred entry, and reduce if the stock gives back the entire post-print move within 1-2 sessions.
• Overweight AMZN vs. the rest of mega-cap tech for a 1-3 month trade: AWS is the cleaner second-order beneficiary if the AI build-out continues, with less single-product risk than NVDA.
• Add a watchlist long in the AI supply chain basket (MU, ANET, TSM) for 3-6 months; these names can outperform if the report is showing real capacity digestion rather than one-off order timing.
• Do not short NVDA outright on this print; the falsifier for any bearish view is another quarter of >80% y/y growth with gross margin holding near current levels and no evidence of customer capex pause.
• If you need a hedge, use a modest QQQ hedge against an NVDA long rather than a naked short; the main risk is not company-specific miss, but a broad AI multiple reset if hyperscaler capex turns out to be front-loaded.