Nvidia Earnings: Is the Central Bank of the AI Revolution Still a Buy After Q3 Results?

Nvidia (NASDAQ: NVDA), the central bank of the artificial intelligence (AI) revolution, just reported Q3 fiscal 2026 results that crushed expectations. Revenue hit $57 billion, up 62% year over year, with data center sales of $51.2 billion beating estimates by $2 billion. CEO Jensen Huang reported that Blackwell graphics processing units (GPUs) are "sold out" for the next 12 months.

For investors, the earnings beat matters less than answering one question: Is AI infrastructure spending sustainable, or are we witnessing the final stages of a bubble?

Image source: Getty Images.

Why Nvidia controls the bottleneck

Every company deploying AI depends on Nvidia's GPUs. These chips handle thousands of calculations simultaneously, making them essential for training AI models and running predictions. Blackwell delivers dramatically faster AI inference workloads -- Nvidia says a full Blackwell NVL72 rack delivers up to 30 times higher large language model throughput at similar power compared to H100-based systems.

The company produces approximately 1,000 server racks per week, with most of Blackwell's capacity through at least 2025 already committed to hyperscalers, including Microsoft, Amazon, and Meta Platforms.

This dominance isn't easily challenged. Advanced Micro Devices claims its MI325X delivers up to 40% better inference throughput than Nvidia's H200 in select workloads. Intel is pushing Gaudi 3 processors. But Nvidia still controls around 90%-plus of the cloud AI GPU market because its CUDA software platform has become the industry standard, creating enormous switching costs.

The spending question

Hyperscaler and AI infrastructure capital expenditures could approach $500 billion to $600 billion by the end of the decade, according to Nvidia and Wall Street estimates. Collectively, Alphabet, Microsoft, Amazon, and Meta are on track to exceed $300 billion in annual capital expenditures, with a growing majority of these directed toward AI infrastructure.

At a $4.5 trillion market cap and about mid-40s times trailing earnings, Nvidia trades at a substantial premium. Furthermore, the bear case draws parallels to Cisco Systems, which dominated the networking equipment market during the late 1990s internet buildout before collapsing by 90% when capital expenditures dried up.

Why this looks more like a transformation

Three factors distinguish today's AI spending from past bubbles, however.

First, applications already generate measurable returns through customer service automation, software development, drug discovery, and autonomous systems. The dot-com era funded business models with no profitability path.

Second, performance keeps improving. Blackwell represents a genuine 30-times leap for inference workloads, with Rubin shipping in 2026. As long as each generation meaningfully expands AI capabilities, demand is expected to grow.

Third, spending comes from companies with fortress balance sheets making multiyear commitments, not speculation.

The counterargument: Those same companies could reverse course if AI monetization disappoints. Nvidia guided for Q4 fiscal 2026 revenue of about $65 billion, representing strong growth but slower than earlier quarters.

The risks

Supply chain bottlenecks create vulnerability. Blackwell chips require Taiwan Semiconductor Manufacturing's advanced CoWoS-L packaging, where capacity is heavily booked into 2026. High-bandwidth memory supply depends on just three vendors: SK hynix, Micron, and Samsung.

Margins show some pressure. Non-GAAP (generally accepted accounting principles) gross margin came in at 73.6%, down from 75% a year ago but recovering from 72.7% in the previous quarter. Management guided Q4 margins back toward 75%.

The Chinese market has essentially vanished. U.S. export rules drove Nvidia's high-end AI GPU business there from roughly 95% market share to almost zero.

The verdict

The investment case hinges on your conviction about the bubble. If AI represents a fundamental technology shift comparable to cloud computing, Nvidia remains the essential play. The supply-demand imbalance and market dominance justify the premium valuation.

However, if AI spending approaches a peak with infrastructure largely in place, the mid-40s times earnings multiple offers a limited safety margin.

The evidence favors transformation. AI workloads continue to expand, new models require exponentially more compute, and applications operate at scale. For long-term investors who accept volatility, Nvidia's market position and technical leadership make it a compelling investment in the infrastructure sector.

George Budwell, PhD has positions in Microsoft, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Cisco Systems, Intel, Meta Platforms, Microsoft, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool recommends the following options: long January 2026 $395 calls on Microsoft, short January 2026 $405 calls on Microsoft, and short November 2025 $21 puts on Intel. The Motley Fool has a disclosure policy.