This website uses cookies

Read our Privacy policy and Terms of use for more information.

KEY POINTS

- Nvidia shares closed at a record $208.27 on Friday, pushing the chipmaker's market capitalization past $5 trillion for the first time since October 2025.

- Hyperscaler commitments totaling more than $650 billion in AI infrastructure spending for 2026 are fueling unprecedented GPU demand and pricing power.

- Traders should watch next week's earnings from Microsoft, Meta, and Amazon for signals on whether capex plans are accelerating or plateauing.

No semiconductor company has ever been worth $5 trillion. Nvidia got there on Friday, closing at $208.27 after a 4.3% surge that capped a 20% rally over the past month and handed the chipmaker a valuation that dwarfs the GDP of most nations. The move came as investors front-ran a wave of hyperscaler earnings this week, betting that Google, Microsoft, Meta, and Amazon will confirm — or expand — capital expenditure plans that already exceed $650 billion collectively for 2026.

The Spending Machine Behind the Rally

The math is straightforward. Every major cloud provider and model developer — from OpenAI and Anthropic to the hyperscalers themselves — is racing to build out data center capacity, and Nvidia's GPUs remain the bottleneck component. TSMC, which fabricates the bulk of Nvidia's chips, reported a 58% jump in Q1 profit just last week, with high-performance computing now accounting for 61% of its revenue. That ratio was below 50% a year ago. The supply chain is telling you the same story the stock price is: AI demand is not slowing.

What makes this cycle different from the dot-com infrastructure boom is that the spending is backed by revenue, not vapor. Microsoft's Azure AI revenue is growing north of 60% year over year. Meta's Llama models are processing billions of daily inference requests across Instagram, WhatsApp, and its advertising stack. Amazon Web Services launched its custom Trainium2 instances last quarter and immediately sold out capacity through mid-2027. These are not speculative bets. They are production workloads that require more compute every quarter.

Nvidia's pricing power reflects that reality. The company's data center segment generated $35.1 billion last quarter, up 73% from the prior year, with gross margins holding above 74%. Competitors are chipping away at the edges — AMD's MI300X has won pockets of inference work, and Google's TPU v6 is handling internal workloads — but none of them have cracked Nvidia's dominance in training, which is where the heaviest spending occurs.

Valuation Tension and the Bull Case

At $5 trillion, Nvidia trades at roughly 38 times forward earnings, a premium that would look absurd for any other chipmaker but arguably makes sense for a company growing revenue at 60%-plus with monopoly-like market share in the fastest-growing technology category in decades. The bear case rests on cyclicality — chip booms always end, and the customers placing massive orders today could pull back if AI monetization disappoints. But the timing of that pullback keeps getting pushed further out as each quarter's results validate the spending.

The broader semiconductor complex is riding Nvidia's wake. The Philadelphia Semiconductor Index is up 18% year to date. Intel's historic 24% single-day surge last Thursday, driven by its own AI data center growth, signals that the rising tide is lifting even the ships that were listing badly six months ago. ASML, Broadcom, and Marvell have all posted double-digit gains in April as investors rotate into anything with AI exposure.

What Comes Next

The real test arrives this week. Microsoft reports Tuesday, Meta and Amazon follow Wednesday and Thursday. The market is not just looking for revenue beats — it is looking for capex guidance. If any of the big three signal that 2026 infrastructure budgets are tracking above plan, Nvidia likely runs further. If even one hyperscaler hints at a pause or efficiency-driven spending reduction, the $5 trillion handle becomes a ceiling rather than a floor. Nvidia's own earnings are due May 28, but the next five days of customer reports will set the tone for the entire AI trade through the summer.

Keep Reading