Nvidia dominates AI chips; AMD wins investor returns in H1 2026

The race to dominate AI chips in the first half of 2026 produced three distinct outcomes depending on what metric you track. Nvidia won on the measure that has traditionally mattered most: actual market dominance and profit generation. Its data center segment alone generated roughly $194 billion（約31兆円） over its most recent fiscal year, a figure that dwarfs AMD's entire data center business more than 11-fold. Nvidia's control of 80–90% of the AI data center GPU market remains essentially unchallenged, and its new Vera Rubin platform has ramped into full production with demand visibility stretching into the trillions of dollars. When judged purely on selling the chips that train and run AI, Nvidia is not just ahead—it is lapping the field. However, Nvidia's stock tells a different story. Shares rose only modestly in the first half despite spectacular results, because expectations were already so high that even strong performance failed to move the needle materially. AMD won the investor-returns battle and grabbed the hearts of Wall Street. Its stock climbed triple digits as the company posted record data center revenue, up 57% from a year earlier. More importantly, AMD secured enormous multiyear commitments from OpenAI and Meta Platforms, deals that provide genuine revenue visibility and validate its position as a rising challenger. AMD remains nowhere near Nvidia's market share—it holds only a mid-single-digit slice of the AI GPU market—but it has clearly established itself as the No. 2 player and delivered the best stock returns of the three. Intel presents a paradox: it is essentially absent from the AI accelerator race, yet its stock soared around 340%, making it the best-performing stock of the three this year. The reason for this disconnect is that Intel's rally reflects investor excitement over its foundry turnaround and a wave of outside backing as it attempts to become a contract manufacturer for others—a bet entirely divorced from AI chips. Intel's Gaudi AI accelerators never gained traction, with Intel itself conceding they would not generate meaningful revenue, and its release schedule has been erratic. Its next-generation data center AI chip is not expected until well into 2027, which means the company has essentially ceded the current AI silicon race to focus on a longer-term foundry strategy. For investors evaluating the three companies, the framework is now clear. Nvidia remains the king with dominant market position and durable profitability, though its stock offers lower upside because expectations are already embedded in the price. AMD is the higher-risk, higher-reward challenger with genuine momentum and credible customers, having delivered the best shareholder returns of the group in the first half. Intel requires an entirely different thesis, one built on foundry prospects rather than AI chips, and appears to be a longer-term bet that may not pay off in the near term.

The first half of 2026 exposed a widening divergence in how the three chip giants are faring in AI. Nvidia's dominance rests on a foundation that has grown harder to challenge: its 80–90% market share in AI data center GPUs is backed by roughly $194 billion（約31兆円） in data center segment revenue—more than 11 times AMD's entire data center business—and its Vera Rubin platform is now in full production with massive demand visibility. Yet despite these commanding fundamentals, Nvidia's stock barely moved because investor expectations were already priced in at such elevated levels. AMD, by contrast, has captured investor imagination by emerging as a credible alternative. Its 57% year-over-year data center revenue growth and marquee multiyear commitments from OpenAI and Meta signal that it is winning real market share and customer confidence. Still, even as the clear No. 2 and a fast-rising one, AMD holds only a mid-single-digit slice of the overall AI GPU market. Intel's story is altogether different: it conceded that its Gaudi AI accelerators would not generate meaningful revenue and is now absent from the active AI accelerator competition. Yet its stock outperformed both rivals, climbing around 340%, because that rally is tied to investor optimism about its foundry business turnaround—a completely separate bet from the AI silicon race itself.

No comments yet. Be the first to share your thoughts!