NVIDIA Q1 FY2027 Earnings: $81.6B Revenue, $91B Q2 Guide — The AI Supercycle Just Raised Its Speed Limit
NVIDIA just delivered what may be the most consequential earnings report in the history of the semiconductor industry. The company reported Q1 FY2027 results after the bell on Tuesday, May 20, blowing past every estimate on the tape: $81.6 billion in revenue, $1.87 in non-GAAP EPS, $2.39 in GAAP EPS, and a Q2 guidance of $91 billion that left Wall Street’s $86.6 billion consensus in the dust. The AI infrastructure supercycle is not slowing down. It is accelerating.
The Headline Numbers
NVIDIA reported Q1 FY2027 revenue of $81.6 billion — an 85% increase year-over-year from $44.1 billion in Q1 FY2026, and a massive beat over the Wall Street consensus of $78.8 billion and even the buyside whisper number of $80 billion. Non-GAAP EPS came in at $1.87, up 140% year-over-year from $0.78, and ahead of the $1.77 consensus. GAAP EPS was even more dramatic at $2.39, up 214% from $0.76 a year ago.
Gross margins expanded significantly — GAAP gross margin hit 74.9%, up 14.4 percentage points from 60.5% in Q1 FY2026. Non-GAAP gross margin was 75.0%, up 14.2 points year-over-year. The margin improvement was driven by lower inventory provisions, primarily because Q1 FY2026 included a $4.5 billion charge associated with H20 excess inventory in China — a headwind that did not repeat. Operating income surged 147% to $53.5 billion on a GAAP basis. Net income exploded 211% to $58.3 billion. These are not software margins — they are the margins of a company with near-total pricing power over the most critical hardware in the global economy.
The Q2 Guide: $91 Billion
The number that will move the stock most is not the Q1 beat — it is the Q2 FY2027 outlook. NVIDIA guided for $91.0 billion in Q2 revenue, plus or minus 2%. Wall Street consensus was $86.6 billion. Goldman Sachs, one of the most aggressive bulls on the Street, had estimated $87.7 billion. NVIDIA’s own guide came in $4.4 billion above the Street consensus — an extraordinary gap that signals the company sees no deceleration in AI infrastructure demand heading into the summer. Non-GAAP gross margin guidance for Q2 is 75.0%, plus or minus 50 basis points, and non-GAAP operating expenses are guided to approximately $8.3 billion.
One critical disclosure in the Q2 outlook: the $91 billion guide does not assume any Data Center compute revenue from China. That means any resolution of U.S.-China chip export restrictions — or any H20/H200 sales approvals — would be pure upside to guidance. Jensen Huang joined Trump’s delegation to Beijing on May 14, and while no chip export deal was announced, the language around China has shifted from closed door to active negotiation. The China optionality embedded in this guide is significant.
Data Center: The Engine
Data Center revenue, NVIDIA’s core business, hit $37.9 billion in the Hyperscale segment and $37.4 billion in AI Clouds, Industrial, and Enterprise (ACIE) — a new reporting structure NVIDIA introduced this quarter to give investors clearer visibility into its two growth vectors. Hyperscale — which includes revenue from public clouds and the world’s largest consumer internet companies — grew from $17.6 billion in Q1 FY2026. ACIE — which captures AI Clouds, Industrial AI, and Enterprise on-premise deployments — grew from $21.5 billion. Together, Data Center drove the overwhelming majority of NVIDIA’s $81.6 billion quarter.
The company noted that Blackwell has been adopted and deployed by every major hyperscaler, every cloud provider, and every model builder. That is not a marketing claim — it is reflected in the revenue. Microsoft raised its 2026 capex forecast to $190 billion. Meta raised its 2026 capex to $125–$145 billion. Amazon, Google, and Oracle are all accelerating AI infrastructure spend. Every dollar of that capex eventually routes through NVIDIA’s order book.
Vera Rubin: The Next Chapter
NVIDIA confirmed that Vera Rubin is on track for the second half of FY2027, starting in Q3. The Vera Rubin NVL72 platform — combining the Vera CPU with Rubin GPUs via NVLink-C2C interconnect — is the architecture that NVIDIA says delivers agentic AI inference at one-tenth the token cost of previous-generation alternatives. Given that NVIDIA hand-delivered the first Vera CPUs to Anthropic, OpenAI, and SpaceXAI just this week, the Q3 ramp is already underway at the customer level.
The Vera Rubin ramp is the single most important near-term catalyst for NVIDIA’s revenue trajectory. The Blackwell-to-Rubin transition is the key watchpoint for gross margins — new architectures typically carry lower initial margins before yields improve at scale. NVIDIA guided Q2 gross margins at 75.0%, suggesting the transition is being managed without margin deterioration. If Rubin follows the Blackwell ramp pattern, the revenue acceleration in Q3 and Q4 FY2027 could push NVIDIA toward a $400 billion annual revenue run rate by the end of the fiscal year.
Free Cash Flow and Capital Return: Record Territory
NVIDIA generated $48.6 billion in free cash flow in Q1 FY2027, up from $26.1 billion in Q1 FY2026 — an 86% year-over-year increase. Operating cash flow was $50.3 billion. The company returned a record approximately $20 billion to shareholders in the quarter through dividends and share repurchases, compared to approximately $14 billion in Q1 FY2026.
The capital return story got dramatically more shareholder-friendly with two announcements: NVIDIA announced a new $80 billion share repurchase authorization, on top of approximately $39 billion remaining as of Q1 FY2027 — giving the company roughly $119 billion in total buyback capacity. It also announced an increase in its quarterly dividend from $0.01 to $0.25 per share — a 2,400% dividend increase that signals management’s confidence in the durability of the cash flow base. For a company that has been primarily a growth story, the emergence of a serious capital return program is a structural shift in the investment thesis that attracts a new category of income-oriented institutional buyer.
The Beat in Context
To put this quarter in perspective: NVIDIA just reported more revenue in a single quarter — $81.6 billion — than Apple typically generates in its best holiday quarter. Its net income of $58.3 billion is larger than the annual revenue of most S&P 500 companies. Its free cash flow of $48.6 billion in a single quarter exceeds what ExxonMobil generates in an entire year. And it guided the next quarter to $91 billion — higher still.
NVIDIA has now beaten revenue estimates for seven consecutive quarters. The Q2 guide of $91 billion represents 69% year-over-year growth on top of a $54 billion Q2 FY2026 base — and it excludes any China revenue. The company’s own commentary on demand is unambiguous: hyperscalers, cloud providers, model builders, sovereign AI customers, and enterprise on-premise deployments are all accelerating simultaneously. Jensen Huang’s statement that NVIDIA expects to sell $1 trillion worth of AI chips across 2026 and 2027 no longer sounds like hyperbole — it looks like math.
What the Market Was Watching — and Got
Heading into tonight’s print, there were four things analysts said would drive the stock reaction: the Q2 revenue guide vs. the $86.6 billion consensus, gross margin trajectory through the Rubin handoff, any language on China and H200/H20 chip sales, and sovereign AI customer concentration. NVIDIA delivered on all four. The Q2 guide of $91 billion cleared the bar by $4.4 billion. Gross margins held at 75.0% through the architecture transition. China was explicitly excluded from the guide — creating upside optionality. And sovereign AI crossed $30 billion in FY2026, more than tripling year-over-year, representing roughly 14% of total revenue and a growing hedge against hyperscaler concentration risk.
NVIDIA also introduced a new capital structure that will attract institutional buyers who previously avoided the stock on the absence of meaningful dividend income. The $0.25 quarterly dividend and $119 billion in buyback capacity create a total shareholder return profile that competes with the best capital allocators in the S&P 500.
The whisper number was $80 billion. NVIDIA printed $81.6 billion. The Q2 whisper was $87–$88 billion. NVIDIA guided $91 billion. The AI infrastructure supercycle just got a new speed limit — and NVIDIA just raised it.
This article is for informational purposes only and does not constitute investment advice. Past performance is not indicative of future results.
About the Author
