KEY POINTS
- Nvidia reports fiscal Q1 2027 results after the close on May 20, with Wall Street consensus at $78 billion in revenue and $1.77 in non-GAAP EPS, both representing roughly 80% and 120% year-over-year growth respectively.
- The four largest hyperscalers have collectively pledged $725 billion in 2026 capex, up 77% from last year's $410 billion, anchoring demand for Nvidia's Blackwell architecture.
- Traders should watch forward guidance closely: consensus Blackwell revenue estimates range from $21 billion to $142 billion for the full fiscal year, making the conference call commentary more important than the print itself.
Nvidia faces the market tonight with a consensus revenue target of $78 billion for its fiscal first quarter ended April 26, and a stock price that has lagged the broader chip sector all year. The AI chip giant reports after the closing bell on May 20, and while the headline number matters, this quarter's real story is about margins, Blackwell shipment cadence, and whether Jensen Huang's guidance validates a capex supercycle that now stretches well into 2027.
The Numbers That Matter
Analysts expect $78 billion in revenue and $1.77 in non-GAAP earnings per share, with data center revenue accounting for roughly $73 billion of the total. Some sell-side desks are modeling closer to $79 billion, and the most aggressive houses have penciled in prints north of $80 billion. The whisper number, as always with Nvidia, runs ahead of the official consensus. History favors a beat: Nvidia has topped estimates in 15 of its last 16 quarterly reports, and the stock has moved an average of 8% in either direction on earnings day over the past two years.
What separates this quarter from prior blowout prints is the margin question. Blackwell now drives the majority of data center compute revenue, and its production ramp has not been seamless. Gross margins contracted slightly in the prior quarter as Nvidia ramped Blackwell production at TSMC, and analysts want to see whether those margins have stabilized or are expanding again. The consensus gross margin estimate sits near 73%, down from the mid-70s Nvidia posted during peak Hopper shipments.
The Hyperscaler Capex Backstop
The demand picture for Nvidia's data center business has never looked stronger on paper. Alphabet, Amazon, Microsoft, and Meta collectively announced roughly $725 billion in 2026 capex plans, a 77% increase from the $410 billion they spent in 2025. That spending underpins GPU procurement at a scale the semiconductor industry has never seen. AWS alone has a chip-related backlog of $225 billion, and its custom silicon business is growing at triple-digit percentage rates year over year.
The question traders are wrestling with is whether this spending benefits Nvidia proportionally or whether it increasingly flows to custom ASIC solutions from the hyperscalers themselves. TrendForce projects 44.6% growth in custom ASIC shipments in 2026 versus 16.1% growth for merchant GPUs, a divergence that did not exist two years ago. Google's TPU v6, Amazon's Trainium3, and Microsoft's Maia 2 are all shipping in volume, and while none replace Nvidia's training monopoly today, they are chipping away at inference workloads that once defaulted to H100s and A100s.
Competition Is No Longer Theoretical
AMD announced it will ramp production of its MI450 GPU in the second half of 2026, and the early demand signals are formidable. AMD has locked in a 6-gigawatt Meta deployment using a custom MI450-based GPU and Helios rack-scale architecture, plus a deal to supply 50,000 MI450 GPUs to Oracle by Q3. Meanwhile, Intel — up more than 225% in 2026 — has re-entered the conversation with its foundry revival and surging Xeon server CPU demand in AI data centers.
None of this dethrones Nvidia. CUDA's software ecosystem remains the deepest moat in semiconductors, and Blackwell's performance per watt on large language model training is still unmatched. But the stock's 15% year-to-date gain against the Nasdaq's comparable return tells you the market is pricing in a maturing competitive landscape, not the monopoly premium Nvidia commanded in 2023 and 2024.
What Comes Next
The conference call at 5:00 PM ET will be the main event. Investors want specifics on Blackwell Ultra timing, supply allocation between hyperscalers and enterprise customers, and any commentary on the China export restriction impact after Washington tightened controls again earlier this year. Revenue guidance for fiscal Q2, expected near $82 billion to $85 billion, will set the tone for whether Nvidia's stock can reclaim the $150 level it last touched in January. If Huang delivers another guide-up and signals margin expansion on Blackwell, the stock could gap higher despite its relative underperformance this year. If the guidance disappoints even modestly, the rotation into Intel, AMD, and Micron accelerates.
