Prediction: AMD Stock Could Outrun Nvidia Over the Next 3 Years -- Here's the Bull Case

The article argues AMD could outperform Nvidia over the next three years as AMD's AI-data-center revenue scales from $6.5 billion in full-year 2023 to $6.7 billion in the latest quarter, up more than 100% year over year. AMD shares have risen 123% year to date versus Nvidia's 23%, though the author flags AMD's elevated forward valuation as a risk. The bullish thesis rests on AMD's smaller revenue base, expanding AI accelerator portfolio and continued multiyear AI-infrastructure demand, while Nvidia's roughly $96 billion quarterly revenue base makes sustained triple-digit growth more difficult.

Analysis

The investable issue is not whether AMD can post faster percentage growth, but whether its accelerator mix can convert into durable gross-margin and software-ecosystem gains. A smaller revenue base makes headline growth optically easy, while NVDA’s installed base, CUDA switching costs, networking integration and customer financing capacity make share displacement materially harder than a single rack-level specification claim implies. Until hyperscaler deployments show repeat orders and improving AMD data-center gross margin, the market should treat performance assertions as marketing rather than proof of a sustained share shift.

A broader AMD ramp is still constructive for the AI hardware supply chain: TSMC, HBM suppliers and rack/power infrastructure providers benefit from a second qualified accelerator vendor even if NVDA retains dominant economics. The second-order risk for NVDA is not immediate revenue loss but a gradual reduction in scarcity pricing; a credible AMD alternative could pressure system ASPs and bundle economics over the next 6-18 months. Conversely, if AMD demand is concentrated in a small number of subsidized or strategic customer wins, inventory digestion could expose the premium multiple quickly.

The near-term signal is weak and largely sentiment-driven, so avoid chasing an outright AMD long after momentum. Over the next 1-3 months, earnings commentary on MI-series backlog conversion, customer concentration, gross-margin trajectory and supply availability matters more than reported data-center revenue growth. The contrarian view is that AMD’s valuation already discounts substantial execution while NVDA’s lower growth rate may be misread as deceleration rather than a consequence of scale; NVDA can outperform if it preserves pricing and raises forward supply or demand visibility.

Key Decisions for Investors

• Initiate a 1:1 dollar-neutral long NVDA / short AMD pair over the next 1-3 months; target 15-20% relative spread compression, with a 10% adverse relative-performance stop. Thesis fails if AMD demonstrates consecutive quarters of accelerator-led gross-margin expansion and broad hyperscaler repeat deployments.
• Do not add outright AMD exposure ahead of the next earnings print. Upgrade to a tactical long only if management discloses backlog conversion, customer diversification and data-center margin evidence sufficient to support earnings revisions rather than revenue-only momentum.
• Maintain or build diversified AI-semiconductor exposure through TSMC as a lower-idiosyncratic beneficiary of multi-vendor accelerator demand over 6-18 months; reassess if leading-edge wafer utilization or customer capex guidance weakens.
• Set an alert around NVDA’s next guidance update: evidence of ASP compression, rising customer qualification of alternative accelerators, or weaker networking attach rates would invalidate the relative-long thesis and favor closing the pair.