Nvidia shares trade at a forward price-to-earnings multiple of 25, slightly below the Nasdaq-100 average of 26 and at a discount to peers like Amazon (28) and Apple (33). This valuation is striking given Nvidia’s 62% year-over-year revenue growth in Q3 2025 and net income surge of 65%. The comparison suggests the market acknowledges Nvidia’s leadership in AI hardware yet remains cautious about future earnings sustainability, making current multiples appear attractive relative to underlying performance.

In Q3, Nvidia reported record revenue of $57 billion, driven by its Data Center segment, where demand for Blackwell GPUs is described by CEO Jensen Huang as “off the charts.” Gross margins held near 70%, and net income reached $31.9 billion. To crystallize value, the company authorized a $62.2 billion stock buyback, which will shave down share count and lift EPS. These figures underscore Nvidia’s ability to convert robust demand into high profitability and return substantial cash to shareholders.

While Nvidia’s model benefits from its role as a pick-and-shovel provider, concerns are rising over hyperscaler and AI startup spending. Goldman Sachs projects hyperscaler capex on AI infrastructure could hit $527 billion in 2026, raising questions about the payback on massive hardware investments. Deutsche Bank warns that OpenAI may burn through $143 billion by 2029. If major customers begin to rein in AI budgets, Nvidia’s lofty gross margins could face downward pressure.

Looking to 2026, Nvidia plans to launch the Rubin class of GPUs tailored for AI video generation, which could open new markets beyond LLM training. Analysts expect continued double-digit data center revenue growth, though share-price gains may moderate as investors weigh product cadence against broader AI spending uncertainty. At a $4.6 trillion market cap, sustaining upside will depend on execution of new GPU architectures and stability in customer capital allocation.