Advanced Micro Devices shares fell as much as 8.8% in premarket trading Wednesday, shaving more than $74 billion from the company’s market value, after a third-quarter revenue forecast that technically beat analyst estimates still failed to satisfy investors seeking unambiguous evidence that the AI spending boom is accelerating AMD’s growth trajectory. The stock’s decline was compounded by SpaceX CEO Elon Musk’s post-earnings disclosure that SpaceX would build its computing infrastructure exclusively on Nvidia chips – describing the Vera Rubin architecture as the best in the industry – a public endorsement that sent Nvidia shares higher by nearly 2% in the same session. FinancialMediaGuide examines this earnings response as a textbook case of the expectations inflation problem that has developed around AI semiconductor stocks, where objectively strong results are punished because the gap between consensus estimates and investor positioning has grown too wide for even a beat to close.
The financial results themselves were not weak by conventional standards. AMD’s data-center revenue more than doubled to $6.72 billion in the second quarter, comfortably exceeding consensus estimates. The company forecast third-quarter revenue of approximately $13 billion plus or minus $300 million, above the $12.52 billion analyst estimate. CEO Lisa Su projected data-center revenue to more than double by 2027 and total revenue growth above the company’s previously outlined target of more than 35%.
The market’s reaction reflects the challenge AMD faces as a growth stock trading at elevated multiples. The company’s shares had more than doubled this year ahead of the earnings release, driven by the thesis that AMD would emerge as the leading alternative to Nvidia in AI accelerators. That positioning creates a compounding expectations problem: when investors have built positions based on AMD closing the gap with Nvidia, any signal that Nvidia is extending its lead – even temporarily – triggers profit-taking disproportionate to the underlying change in AMD’s financial trajectory. Bernstein analyst Stacy Rasgon noted that expectations had moved higher following Intel’s results and that the buyside already had a fairly bullish outlook. FinancialMediaGuide highlights this positioning dynamic as the most structurally important context for interpreting AMD’s share price decline, since it explains why a quarter with doubled data-center revenue and above-consensus guidance can still produce an 8.8% negative session.
The SpaceX disclosure deserves particular attention because it was made publicly and specifically by name. Musk’s characterisation of Nvidia’s Vera Rubin as the best architecture is the kind of high-profile customer endorsement that carries extraordinary weight in the AI chip market, where a handful of dominant spenders – hyperscalers, AI frontier labs, and now SpaceX – collectively determine the commercial trajectory of every semiconductor company in the ecosystem. When the world’s most watched technology entrepreneur publicly selects a competitor’s architecture, the market reads it as a data point in the competitive assessment regardless of whether AMD was ever seriously under consideration for SpaceX’s specific use case.
AMD’s recent deal activity suggests its competitive positioning is not as compromised as the single-day stock move implies. Last month the company announced partnerships with Anthropic and Core Scientific specifically to bolster AI infrastructure ambitions – two customers with significant credibility in the frontier AI and data center markets respectively. Those relationships indicate that AMD’s MI series of AI accelerators is finding meaningful enterprise adoption, even if the largest training workloads at hyperscalers remain dominated by Nvidia. TD Cowen analysts acknowledged this tension, describing the AMD quarter as facing a very high bar following recent AI customer announcements and a sharp rally, despite objectively good results and guidance. FinancialMediaGuide notes that the Cowen characterisation – objectively good but failing a very high bar – captures precisely the predicament that elevated valuations create for any company in a sector where investor expectations have disconnected from financial fundamentals.
The competitive dynamic with Nvidia has structural dimensions that a single quarter’s results cannot resolve. Nvidia’s CUDA software ecosystem represents a switching cost that AMD’s ROCm platform has not yet eroded to a degree that meaningfully shifts large-scale model training workflows. The Vera Rubin architecture that Musk endorsed specifically is Nvidia’s next-generation accelerator platform, designed for the agentic AI and inference-at-scale workloads that are increasingly defining the frontier of AI system design. AMD’s ability to compete in those workloads – as opposed to the training-intensive applications where it has made more demonstrable progress – is the key commercial question for the second half of 2026.
AMD’s data-center GPU revenue trajectory is nonetheless remarkable in context. Doubling data-center revenue in a single quarter is a commercial achievement that most semiconductor companies would regard as a major positive. The issue is not that AMD is failing; it is that it is succeeding at a pace that was already assumed by investors who bid the stock to more than double its year-ago level. The next phase of AMD’s AI story will require either a major customer win that demonstrably shifts the competitive landscape with Nvidia, or a sustained period of financial execution that compounds the current growth trajectory into revenue scales that justify the current multiple through earned rather than expected earnings power.
For investors assessing AMD at current prices, the core question is whether the 8.8% decline has reset expectations to a level where the actual financial trajectory can now produce positive surprises, or whether the stock is still priced for a competitive outcome that requires Nvidia to stumble in ways that are not visible in either company’s current commercial pipeline. Lisa Su’s stated targets – data-center revenue more than doubling by 2027, total growth above 35% – give investors a specific measurement benchmark against which quarterly execution can be tracked, and Financial Media Guide identifies the next two quarterly earnings calls as the period in which the market will determine whether AMD’s AI infrastructure positioning is a durable competitive force or a credible but ultimately secondary alternative to Nvidia’s dominant franchise.
