NVDA Q1 Revenue Beats at $81.6B but Stock Falls 2.5% After Hours
A 140% YoY EPS beat and $91B Q2 guidance failed to lift NVDA as expectations had already priced in the raise, exposing how high the bar has become. With 30-year yields at 17-year highs, the cost-of-capital math is pressuring hyperscaler AI capex multiples across ^IXIC.
RKey facts
- Nvidia Q1 revenue $81.6B, up 85% YoY vs $74B expected; EPS $1.87 (+140% YoY) vs $1.78 expected
- Q2 guidanceCompany-issued forecasts of future financial performance. $91B excludes China data center revenue; guidance raises priced in, stock fell 2.5% AH
- Amazon to add 1M Blackwell/Rubin GPUs this year, worth $30-40B in chip spend at 75 GPUs per rack
- US 30-year yield at highest level since 2007; markets price 37% odds of Fed rate hike in 2026
- H100 (three-generation-old GPU) rental prices up ~20% in 2026 despite age
What's happening
Nvidia delivered blockbuster Q1 results that should have sent shares soaring, yet the stock stumbled in after-hours trading. Revenue hit $81.6 billion against $74 billion expectations, while data center revenue alone reached $75.2 billion, nearly doubling year-over-year. Earnings per share came in at $1.87, a 140% year-over-year jump versus consensus of $1.78. The company also announced an $80 billion share buybackA company repurchasing its own shares from the open market. program and guided Q2 revenue to $91 billion, materially above Street consensus of $84-86 billion.
The disconnect between the beat and stock weakness reveals a crucial dynamic: expectations had reached such lofty levels that even extraordinary results failed to justify current valuations. Sell-side analysts were nearly uniformly bullish heading into the print, and retail options positioning skewed heavily bullish. A clean beat with flat guidanceCompany-issued forecasts of future financial performance. would have likely triggered profit-taking, but the raise appears to have been priced in already. Investors are now questioning whether the bar for future beats will only climb higher, and whether AI infrastructure capex, now running at scale across major hyperscalers, can continue absorbing the rising cost of capital.
The real debate centers on sustainability. Amazon disclosed it will add 1 million Blackwell and Rubin GPUs this year alone, worth $30-40 billion in chip spend. Google and Meta are deploying hundreds of billions into data center infrastructure. But with US 30-year yields at their highest since 2007, and markets pricing in 37% odds of a Fed rate hike in 2026, the calculus is shifting. Hyperscalers must prove their AI investments are generating measurable ROI, not just capacity for its own sake.
Skeptics note that H100 rental prices (a three-generation-old GPU) have risen 20% in 2026, suggesting constrained supply and strong near-term demand. But Nvidia's exclusion of China data center revenue from its guidanceCompany-issued forecasts of future financial performance. is a candid acknowledgment of geopolitical risk. If Beijing reopens its spigot, or if US export controls tighten further, the growth narrative could face a sharp reset. For now, the market is punishing the stock despite the beat, betting that hyperscaler capex discipline will eventually catch up with reality.
What to watch next
- 01Amazon AWS capex updates and hyperscaler earnings: next 4 weeks
- 02US 30-year bond yield trajectory: ongoing
- 03Fed rate path and geopolitical resolution on China controls: June FOMCThe Federal Open Market Committee - the Fed's rate-setting body.
- CNBC Top NewsMeta settles first U.S. case over school costs tied to youth mental health, court filing shows
The agreement fully resolves a lawsuit brought by a Kentucky school district following earlier settlements by co-defendants YouTube, Snap and TikTok.
40m ago - CNBC Top NewsAnthropic, Microsoft in talks for AI chip deal after $5 billion investment
Microsoft has not made the Maia 200 chips available to customers, but it is used in the company's data centers, offering better efficiency than other silicon.
41m ago - Yahoo FinanceNVIDIA (NVDA) Posts Blowout Q1 on AI Demand, Forecasts Another Record Quarter57m ago
- Yahoo FinanceMeta settles first US case over school costs tied to youth mental health, court filing shows58m ago
- CNBC Top NewsAn AI trade involving energy and infrastructure that's doubled your money, topping Nvidia
If you put the same money into a basket of companies that are building out AI infrastructure and energy sources, you’ve done much better than stocks like Nvidia.
1h ago - Yahoo FinanceNvidia Beats, Stock Dumps—BofA Says Buy the Dip1h ago
- Yahoo FinanceJensen Huang Said Something Surprising About AI. Here's Why Nvidia Investors Should Pay Attention.1h ago
- BloombergOpen Interest 5/21/2026
Get a jump start on the US trading day with Matt Miller and Dani Burger on "Bloomberg Open Interest." Nvidia’s blowout earnings fail to supercharge the AI trade, as investors eye the next wave of mega IPOs from SpaceX, OpenAI, and Anthropic. Plus, Jamie Dimon’s AI hiring shift: more AI staff, fewer bankers Plus, warnings on the AI debt boom, e.l.f Beauty’s Rhode-fueled surge, and a Trump advisor’s prediction for lower gas prices. (Source: Bloomberg)
1h ago
Related coverage
- Top 10 S&P 500 Names at 38% Weight While Russell 2000 Trails Mag-7 by 40 Points in 2025Equities US··0 mentions
- NVDA Drops 2.5% After Hours Despite 85% Revenue Growth and $91B GuidanceTech & AI··0 mentions
- NVDA Q2 Guidance of $91B Beats Estimates as Stock Dips 2.5% After HoursTech & AI··0 mentions
- Top 10 S&P 500 Stocks at 38% Index Weight, Highest Concentration Since 2021Equities US··0 mentions
More about $NVDA
- META's $145B Capex Plan Accompanies a 10% Workforce Cut Toward AI Redeployment·Tech & AI
- SpaceX IPO Discloses 18,712 BTC and a $26.5T AI Addressable Market Target·Tech & AI
- NVDA Q2 Guidance of $91B Meets a 30-Year Treasury Yield at 2007 Highs·Tech & AI
- META Cuts 8,000 Roles While Holding $145B Capex Guidance Near $614 Stock Price·Tech & AI
- NVDA Guides 94% Q2 Growth but Slides 2.5% as China Exclusion Clouds $91B Target·Tech & AI
Tracking AI infrastructure capex — hyperscaler spend, data center buildouts, memory demand and the margin compression risk.