This AI Giant Has Climbed 900% Over the Past Five Years. Now Wall Street Expects the Stock to Jump Another 40%.

Nvidia says data center revenue rose 90%+ to a record $75B in the quarter, supporting $81B total revenue and gross margins above 70%. The article highlights expectations for another ~40% stock upside over the next 12 months, while outlining a new growth push into CPUs with the Vera Rubin platform and guidance for $20B in stand-alone CPU revenue this year. It also notes a recent cooling (shares up ~7% H1) from AI capex concerns and chip competition risk from hyperscalers.

Analysis

The incremental profit pool is not the headline GPU unit volume; it is the ability to sell a fuller rack-level stack and make customers pay for integration, not just compute. That strengthens NVDA’s pricing power and makes its revenue less cyclical than a pure chip vendor, while raising the hurdle for AMD and especially INTC to defend sockets in data-center workloads. The second-order loser is the custom-silicon narrative at AMZN and META: every move to internal chips is a hedge against supplier concentration, but if NVDA becomes the default control plane for AI infrastructure, those designs risk being confined to edge cases rather than broad replacement.

Near term, the stock’s path is more about conviction on 2025-26 demand than current-quarter beats. The market will likely fade the story unless management can show that CPU attach rates are real, gross margins stay structurally above peers, and hyperscaler capex translates into visible backlog rather than just forward-looking rhetoric. The key falsifier is any sign that AI system spending is being delayed, re-specified toward lower-cost custom silicon, or that new platform launches fail to convert into enterprise orders within 1-2 quarters.

Contrarian view: consensus may be underpricing how much NVDA can expand its addressable wallet share even if it loses share on the GPU leg, but it may also be overpricing the speed of monetization in CPUs. A realistic base case is that this becomes a margin-defense story first and a share-gain story second; if so, upside is steadier but less explosive than the most bullish narratives imply. For INTC, the risk is not total displacement but continued inability to show data-center CPU stabilization, which would keep multiple compression in place for 6-18 months.

Key Decisions for Investors

• Maintain a tactical long NVDA on pullbacks, but size it as a quality compounder rather than a momentum trade; best entry is after any post-earnings/launch air pocket, with a 3-6 month horizon and downside tied to any guide-down in AI capex or margin.
• Pair trade: long NVDA / short INTC for 1-3 months into the next platform cycle. Thesis is that INTC bears the cleanest competitive displacement risk in data-center CPUs; invalidate if INTC shows sustained data-center share stabilization or positive guide revisions.
• Avoid chasing AMD on the CPU AI narrative until there is evidence that NVDA’s standalone CPU is gaining enterprise design wins; if you want exposure, prefer NVDA over AMD on relative valuation and system-level integration.
• Watch AMZN and META as bellwethers for custom-silicon substitution. If either starts signaling broader internal-chip deployment beyond internal workloads, reassess NVDA multiple expansion; if not, the market is likely overestimating the pace of share erosion.
• Set an alert on NVDA valuation/margin compression: if forward P/E expands while gross margin or data-center growth decelerates for two consecutive quarters, take profits and rotate into less consensus AI infrastructure names.