AMD Stock and Intel Crushed Nvidia in the First Half. Here's My Prediction for the Second Half.

The article argues that Nvidia can regain relative leadership in the second half as it targets the $200B data-center CPU market with its first stand-alone CPU (Vera Rubin) plus a fall “superchip” combining GPU+CPU. It cites Nvidia forecasting $20B in stand-alone CPU sales this year, supported by an expected shift toward agentic AI, and contrasts Nvidia’s valuation (~22x forward earnings) as “dirt cheap” versus AMD/Intel after their first-half surges (+171% for AMD and +278% for Intel). Overall, the piece expects Nvidia’s stock to outperform rivals despite a recent momentum pause.

Analysis

The market is still pricing AI as a component-volume story, but the real shift is platform control: if one vendor owns the accelerator, CPU, networking, and software stack, it can capture a larger share of every data-center dollar and make it harder for rivals to win on standalone silicon. That is most negative for AMD, whose recent rerating assumes continued share gains in server compute, and secondarily for INTC, where any loss of relevance in AI-adjacent server sockets slows the turnaround narrative and keeps the multiple capped.

The near-term trade is less about unit shipments than validation. Over the next 1-3 months, watch for channel checks, hyperscaler qualification timelines, and whether the fall product cycle is presented as a full-rack economics upgrade rather than a chip launch; that would support a faster-than-feared mix shift into higher-margin platform attach. If adoption slips into 2026 or customers keep treating CPUs as commoditized plumbing, the stock reaction should fade quickly because the market will not wait for multi-year TAM slides to close.

Contrarian view: the consensus may be underestimating how much of the upside is already embedded in NVDA’s AI franchise, while overestimating how easy it is for peers to monetize their catch-up rally. The more important question is not whether NVDA can win some CPU sockets, but whether it can raise gross margin dollars per deployed rack; if yes, the valuation gap can remain justified even after the stock’s run. The thesis is falsified if the fall launch fails to show material design wins, if AMD sustains server-share momentum through the next two earnings prints, or if INTC’s data-center revenue inflects more strongly than expected.

Key Decisions for Investors

• Long NVDA / short AMD on a 2-4 month horizon into the fall product cycle; use the pair to isolate platform-share upside from broader AI beta. Risk/reward favors NVDA if the market starts discounting CPU attach as incremental margin, but cover if AMD continues taking server share or if NVDA’s launch commentary is qualitative only.
• Avoid chasing INTC on the long side here; the stock has a turnaround narrative and policy support, but that also makes it a poor short candidate. If you want a cleaner relative-value expression, keep INTC as a hedge leg only when paired against NVDA strength rather than as a standalone bearish outright.
• Buy NVDA on post-event or post-earnings pullbacks rather than pre-launch momentum. The entry should be conditioned on confirmation that management quantifies CPU-related revenue or platform attach; without that, the move risks becoming a sentiment trade with poor follow-through.
• Watch AMD’s next server-CPU commentary as the falsifier for the NVDA outperformance thesis. If AMD maintains or expands data-center share despite NVDA’s CPU push, rotate out of the pair quickly; that would suggest the addressable opportunity is smaller or slower than the market is assuming.
• If you need a lower-conviction expression, consider a modest long NVDA position versus SOXX rather than a single-name short. That captures the idea that this is an index-weight leadership story, while limiting idiosyncratic risk if the broader semiconductor tape rolls over.