Nvidia reports revenue surge and forecasts growth

NVDA----BTC----

Nvidia projected $108 billion in fiscal third-quarter revenue after reporting another sharp acceleration in sales of AI data center equipment, signaling that demand for advanced computing infrastructure remains strong even without expected revenue from mainland China data center customers.

The company said on Aug. 26 that fiscal 2027 second-quarter revenue reached $96.221 billion for the period ended July 26, up 106% from a year earlier and 18% from the preceding quarter. Nvidia’s forecast for the current quarter, allowing for a 2% variation in either direction, would extend that growth streak while excluding mainland China data center computing revenue from its outlook.

Data center products accounted for $89.023 billion of second-quarter revenue, according to Nvidia, or more than 92% of total sales. The division grew 117% year over year and 18% sequentially, making it the dominant source of the company’s expansion from the prior quarter’s $81.615 billion in total revenue.

Nvidia’s numbers reinforce the scale of capital being directed toward AI computing capacity. They do not, by themselves, establish a direct pricing catalyst for Bitcoin or AI-linked crypto tokens, whose valuations depend on token-specific supply, network activity, liquidity and broader financial conditions rather than on chip sales alone.

Profit margins remain elevated as AI systems scale

Nvidia reported net profit of $59.688 billion, up 126% from the same quarter a year earlier, while diluted earnings per share rose 128% to $2.46. Operating profit increased 124% to $63.734 billion.

On a non-GAAP basis, Nvidia posted net profit of $53.954 billion, an increase of 118% year over year. Its non-GAAP gross margin reached 75%, compared with 72.4% in the year-ago period and 74.9% in the prior quarter.

The margin performance suggests Nvidia has retained substantial pricing power while customers build increasingly large clusters of graphics processing units, networking equipment and related software. Revenue growth alone can sometimes reflect higher component costs or lower-margin system sales; Nvidia’s 75% gross margin indicates that, during the quarter, its product mix and pricing continued to support profitability.

Shares initially fell about 1.3% in after-hours trading after the results were released, before rising more than 4%, according to the supplied market data. The reaction reflected a familiar dynamic around Nvidia earnings: traders are measuring extraordinary reported growth against similarly extraordinary expectations for future AI spending.

Cloud and enterprise demand outpaced hyperscale growth

Nvidia’s revised data center disclosure split the business between hyperscale customers and a category covering AI cloud providers, industrial groups, enterprises, sovereign AI programs and AI-native companies.

Hyperscale customers generated $48.71 billion in quarterly revenue, up 102% from a year earlier and 13% from the first quarter, Nvidia said. The AI cloud, industrial and enterprise category brought in $40.313 billion, rising 138% year over year and 25% sequentially.

That faster growth in the second category points to expanding demand outside the largest established cloud platforms. AI cloud providers rent computing capacity to customers, while sovereign AI programs involve national or regional efforts to build domestic computing resources. Both can create new sources of demand for Nvidia systems, though their spending patterns may be less predictable than those of the largest technology companies.

Nvidia said Blackwell Ultra deployments expanded during the quarter and drove data center growth. It also said its next-generation Vera Rubin platform had entered full production, with rack-scale systems running on partner cloud platforms including CoreWeave and Google Cloud.

The company separately said it had placed Groq 3 LPX, designed for interactive AI inference, into full production. Inference refers to the stage when a trained AI model generates responses or performs tasks. Nvidia also introduced DSX, a platform intended to combine computing, networking, software and systems for large AI data center operations.

China contribution remains minimal in reported sales

Sales of Hopper data center products shipped to mainland China represented less than 1% of data center revenue during the quarter, Nvidia said. The company’s third-quarter guidance includes no mainland China data center computing revenue.

The exclusion puts greater weight on demand from cloud providers, enterprises and government-backed computing projects elsewhere. Nvidia’s $108 billion forecast suggests management expects those markets to compensate for the lack of anticipated China data center sales in the quarter ahead.

Outside its main data center segment, Nvidia reported $7.198 billion in edge computing revenue, up 27% year over year and 13% sequentially. The company attributed the increase partly to Blackwell workstation sales, while stating that higher memory and system prices pressured consumer PC demand.

Commitments show the scale of the infrastructure buildout

As of July 26, Nvidia reported $360 billion in future purchase and investment commitments. The total included $279 billion in supply and capacity commitments, $29 billion in cloud service agreements, $23 billion in capital expenditures and $25 billion in equity investments.

Those obligations show that Nvidia’s current sales growth is tied to a much larger industrial expansion involving chip supply, power availability, cloud capacity and data center construction. The company ended the quarter with $56.6 billion in cash, cash equivalents and marketable debt securities, after issuing $25 billion of senior unsecured notes for general corporate purposes.

Nvidia also described credit support tied to SB Energy’s PORTS-Pike project in Ohio. The initial build is expected to involve roughly 4.25 gigawatts of land, power and facilities intended to host OpenAI infrastructure using Nvidia systems. Nvidia reported maximum guarantee obligations of as much as $105 billion, structured to take effect in stages after conditions including service-ready data centers are met.

The company said it had formed strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to seek more than $500 billion in third-party capital for AI infrastructure. It returned roughly $26 billion to shareholders through repurchases and dividends during the quarter, leaving about $99 billion under its repurchase authorization.

For cryptocurrency markets, the report offers a clear measure of demand for AI hardware but limited evidence for treating Nvidia’s earnings as a standalone trading signal. Tokens connected to computing networks may benefit only where their protocols generate verifiable usage, require tokens for services, or capture revenue from workloads. Nvidia’s results instead place the immediate financial gains from the AI buildout with the hardware supplier and the operators financing physical data center capacity.