AI Trends

NVIDIA's Q3 revenue of $115.8 billion: first earnings report after Jensen Huang's departure beats expectations

Jiwei.comUpdated Nov 18 14:12

AI summary

NVIDIA's revenue for the third quarter of fiscal 2027 was $115.844 billion, up 103.1% year-over-year, while data center revenue was $107.6 billion, up 110% year-over-year, both exceeding market expectations. The company expects fourth-quarter revenue of $126.1 billion to $133.9 billion and has raised its fiscal 2028 revenue growth forecast to 75%. Board Chairman Harvey Jones said there are no plans to reinstate Jensen Huang, that the CEO search is proceeding both internally and externally with no timeline, and that the product roadmaps for Vera Rubin Ultra and Feynman remain unchanged.

Why read this

SelectedAI score 79

The earnings report provides data on revenue, order backlog, and price increases, offering insight into the strength of AI computing demand and the strategic continuity in the post-Jensen Huang era.

Topics

Tags

Body · AI translation

Viewpoint 1: Nvidia's third-quarter revenue was $115.8 billion, up 103% year over year, and data center revenue was $107.6 billion, up 110% year over year. All metrics comprehensively beat market expectations, marking the 17th consecutive quarter of beating Wall Street revenue expectations, with no sign of slowing demand for AI compute.

Viewpoint 2: Facing the market's reassessment of the sustainability of massive AI spending, Nvidia raised its fiscal 2028 revenue growth forecast to 75%, cloud business backlog rose to $2.5 trillion, Hopper re-contracted prices rose by up to double, and B300 prices increased by more than 60%, giving a positive answer on the authenticity of demand.

Viewpoint 3: The first earnings report after Jensen Huang's resignation landed smoothly: the board made clear it would not arrange for his return, the CEO search is proceeding both internally and externally with no timetable; Vera Rubin Ultra will launch next year, with Feynman taking over in 2028, and the compute expansion strategy and product roadmap remain unchanged.

I. Quarterly revenue exceeds $100 billion for the first time: data centers carry the load alone

The earnings report shows that in the third quarter of fiscal 2027 (ended October 25), Nvidia achieved revenue of $115.844 billion, up 103.1% from $57.006 billion in the same period last year and up 20.3% quarter over quarter from $96.221 billion, beating market expectations of $109.33 billion and also clearly above the midpoint of the company's previously provided guidance of $108 billion, marking the 17th consecutive quarter of exceeding Wall Street revenue expectations. Quarterly revenue stood above the $100 billion mark for the first time, GAAP net profit reached $65.287 billion, up 105% year over year, and diluted earnings per share were $2.70.

In terms of earnings quality, non-GAAP operating profit was $77.88 billion, up 116% year over year; non-GAAP net profit was $64.87 billion, up 109% year over year and up 20% quarter over quarter; adjusted earnings per share were $2.68, up 111% year over year and up 21% quarter over quarter; adjusted gross margin was 74.9%, above market expectations of about 74.1%.

By business segment, data center revenue was $107.6 billion, up 110% year over year and up 21% quarter over quarter, accounting for about 93% of the company's total revenue and exceeding market expectations of $101.87 billion; edge computing and other businesses generated revenue of $8.2 billion, up 42% year over year and up 14% quarter over quarter. Product progress was the biggest highlight of the quarter: the new-generation Vera Rubin NVL72 computing platform has entered volume ramp-up and begun deliveries, setting the company's fastest product capacity ramp record in history. Related systems are already running at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud and Nebius, and the company is also deepening cooperation with Meta, OpenAI, Anthropic and others to meet growing computing demand for AI model training, inference and agent applications.

[Chart: Year-over-year growth rates of Nvidia's main metrics in the third quarter of fiscal 2027]

II. Whether it can surprise the market again: expectations are full, and the outlook becomes the deciding factor

Over the past four quarters, Nvidia's earnings all exceeded market expectations, and "beating expectations" has become almost the norm for the market. Before this earnings report, the company's guidance for the third quarter was $108 billion, plus or minus 2% (about $105.8 billion to $110.2 billion), while Wall Street consensus expectations were concentrated around $109 billion. Analysts surveyed by FactSet expected an average of about $108.9 billion, and the LSEG measure was about $109.2 billion. Analysts had already set expectations quite full, leaving limited room for surprise. Combined with the management turmoil triggered by Jensen Huang's resignation, Nvidia's stock price had fallen more than 30% from its October high before the earnings release, and market sentiment leaned cautious.

The final result was still satisfactory: actual revenue of $115.8 billion exceeded the upper end of guidance by about 5%, and adjusted earnings per share of $2.68 were also above the market consensus of about $2.48. But for the stock price, the outlook is more important than ever. Nvidia expects fourth-quarter revenue of $126.1 billion to $133.9 billion (midpoint about $130 billion, plus or minus 3%), exceeding market expectations of $125.25 billion; the guidance does not include data center computing revenue from China, and the fourth quarter includes an extra week. Management also raised its full-year fiscal 2028 outlook, expecting sales to grow 75% year over year, above the previous expectation of 70%.

[Chart: Nvidia quarterly revenue and fourth-quarter guidance trend]

What must be acknowledged is that the quality of the surprise is being diluted by profit margins. The company expects fourth-quarter GAAP and non-GAAP gross margins between 71.0% and 72.0% (midpoint about 71.5%), below 74.9% in the third quarter; Kress said the fourth quarter will be the low point for gross margin, then recover to the 72% to 73% range in fiscal 2028 (midpoint about 72.5%).

From price signals, according to Kress, under multi-year agreements, re-contracted prices for Hopper GPUs reached up to twice the original price, while B300 prices have risen more than 60% since launch. Against the backdrop of compute scarcity, customers are willing to pay a premium for existing capacity, which is the most direct footnote to demand strength. From the order backlog, Nvidia's cloud business backlog reached $2.5 trillion, up from $2 trillion three months earlier. From supply commitments, to lock in capacity, Nvidia's supply chain prepurchases and capacity commitments increased from $119 billion in the first quarter of fiscal 2027 to $279 billion in the second quarter, with the main increase directed toward memory chips; the company has also previously acknowledged that overall supply chain capacity can currently meet only about 70% of customer demand.

[Chart: Path of changes in Nvidia's adjusted gross margin]

Of course, the costs are also real: the extreme pricing environment for memory chips is eroding profit margins, and the fourth-quarter gross margin guidance of 71.0% to 72.0% means that even with higher revenue next quarter, the gross profit left from each dollar of sales will be thinner. The company has also tightened the terms of supply chain financing: it will "strictly review" the recipients of financing support, data center guarantees will start only after facilities are ready and will not be provided during construction; if data center construction is delayed, the guarantee will no longer be valid; in the future, it will be more "disciplined and selective" regarding investments, commitments and guarantees, evaluating them project by project.

Overall, although the challenges facing Nvidia are still accumulating, this earnings report shows that the company will continue to benefit from the surge in demand for artificial intelligence compute.

IV. The post-Jensen Huang era: successor and strategy

On questions of market concern such as how the successor will be chosen and whether the established compute expansion strategy can continue, this earnings report and conference call provided some clues.

Board Chairman Harvey Jones explicitly stated on the conference call that the company does not intend to bring Jensen Huang back and that the board will consider both internal and external candidates. "We want to appoint a permanent chief executive as soon as possible, but we will not rush the decision just to save time; we must first find the most suitable person." According to reports, Nvidia has hired executive search firm Heidrick & Struggles to lead the CEO search.

On the continuation of the established strategy, Kress reiterated that Vera Rubin has had a strong start and the product roadmap is proceeding as planned: the company is still on track to launch Vera Rubin Ultra next year and Feynman in 2028. The raised fiscal 2028 growth outlook of 75% can, to some extent, be seen as the interim management's endorsement of the established compute expansion strategy.

Source

Jiwei.com