Nvidia's last quarter revenue doubled and exceeded expectations; this quarter is expected to surpass the $100 billion mark for the first time, with next fiscal year's growth guidance crushing estimates | Earnings Watch

Nvidia, the global leader of the AI industry chain, saw its revenue accelerate last quarter, setting another quarterly record. The company expects this quarter’s revenue to surpass the $100 billion mark for the first time in history. Although the gross margin guidance for this quarter is slightly below expectations, the revenue growth guidance for the next fiscal year crushes estimates, temporarily reversing market sentiment.

On Wednesday, June 26 (US Eastern Time), after the close of trading, Nvidia released its earnings report for the second quarter of its fiscal year 2027 (ending July 26, 2026). Revenue rose 106% year-over-year to $96.221 billion, exceeding analyst expectations by more than 4%. Non-GAAP adjusted earnings per share (EPS) increased 120% year-over-year to $2.22, nearly 6% higher than consensus estimates.

AI data centers remain the absolute main engine. In the second quarter, data center revenue grew 117% year-over-year to $89 billion, nearly 4% above analyst expectations and accounting for over 90% of Nvidia’s total revenue. Compared to the first quarter's growth rate, both Nvidia’s total revenue and data center revenue accelerated. Actual Q2 revenue exceeded the company’s previous guidance midpoint by over $5 billion.

Looking ahead to the third quarter, Nvidia expects median revenue of $108 billion, nearly 90% year-over-year growth, higher than analyst estimates of $105.15 billion. This means Nvidia will likely officially enter the era of single-quarter $100 billion+ revenue. The company also expects a non-GAAP gross margin midpoint of 74.0% for Q3, slightly below the estimated 75%, and non-GAAP operating expenses to rise to about $9 billion.

After the earnings release, Nvidia’s stock price initially rose slightly in after-hours trading but then quickly turned lower, falling as much as 4% after hours. Analysts believe the after-hours decline does not mean the results are weak or that the Nvidia-AI story is over. Gross margin guidance for Q3 was slightly below expectations, revenue guidance did not exceed the most optimistic Wall Street range, and the guidance excludes China’s data center computing revenue, making it more conservative and leaving China’s potential recovery as an uncertainty for now.

During the subsequent earnings call, Nvidia management noted that demand is still accelerating at the current scale, expecting revenue to grow about 70% in fiscal year 2028, significantly above analysts’ estimate of around 45%. Amazon also committed to greatly increasing the adoption of Nvidia products, deploying an additional 2 million GPUs. Nvidia’s stock rebounded after hours and was up nearly 5% at one point.

Before the results, Nvidia shares had already risen more than 10% this month, reflecting many optimistic expectations. For such a core AI asset, the market is trading on “whether it can greatly outperform the buy-side’s expectations” rather than simply surpassing consensus estimates.

In Q2, Nvidia posted revenue of $96.221 billion, a 106% year-over-year increase—doubling and marking the fastest growth since Q2 FY2025 (ending July 2024). Quarter-over-quarter, revenue rose 18%, showing continued high-speed expansion in AI infrastructure demand. The Q2 revenue guidance midpoint given in Q1 was $91 billion; actual results beat this by about $5.2 billion.

Profitability was also strong. On a non-GAAP basis, Q2 adjusted net profit was $53.954 billion, up 118% year-over-year and 18% quarter-over-quarter. EPS was $2.22, above market expectations of $2.09. Non-GAAP operating profit stood at $63.956 billion, up 124% year-over-year and above the analyst estimate of $61.19 billion, equating to an operating margin of about 66%, still extremely high.

For gross margin, both GAAP and non-GAAP were 75.0% in Q2, basically matching expectations and up 2.5 to 2.6 percentage points from a year prior. For a company approaching $100 billion in quarterly revenue, maintaining a ~75% gross margin shows Nvidia’s strong pricing power in AI accelerated computing markets.

Additionally, adjusted Q2 operating expenses were $8.232 billion, lower than expectations of $8.32 billion.

These numbers mean Nvidia not only beat revenue expectations again but also better controlled expenses than expected.

It’s worth noting that Nvidia’s Q2 “other income” includes about $7.77 billion net gain from equity securities, which is excluded from non-GAAP profit. GAAP EPS was $2.46, higher than the non-GAAP $2.22, mainly because GAAP includes this nearly $7.8 billion gain. Excluding this investment income, non-GAAP net profit was $53.954 billion, better reflecting core operating profitability.

Q2 data center business revenue hit $89 billion, up 117% year-over-year and 18% quarter-over-quarter, higher than analyst expectations of about $85.8 to $85.9 billion. This segment contributed about 92.5% of overall revenue, meaning Nvidia is now tightly tied to the global AI infrastructure build-out cycle.

Nvidia CEO Jensen Huang emphasized in the report: “AI is at a tipping point. It is doing useful work, tokens are becoming productive and profitable. Now, compute equals revenue.”

He also said that a year ago, a single AI laboratory was driving construction, while now, multiple cutting-edge labs are scaling up in parallel, open-source model ecosystems are thriving, and physical AI systems are coming online.

The report discloses that the Vera Rubin platform is in full mass production, with racks operating at partners including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, and Nebius. Nvidia’s Spectrum-6 switch system, Vera CPU, DSX platform, and others are also presented within the “AI factory” narrative, showing Nvidia is shifting from just selling single GPUs toward system-level deliveries, including full racks, networks, CPUs, software, and deployment solutions.

Within data centers, the most watched segment—hyperscale revenue—was $48.71 billion, beating the $43.55 billion estimate, indicating major cloud providers and hyperscale customers remain the AI infrastructure investment main force. Compute & Networking revenue was $88.3 billion, also above the $84.69 billion estimate.

However, not all sub-segments beat expectations. ACIE revenue (including AI cloud, industrial, and enterprise apps) was $40.31 billion, below the $41.96 billion market forecast. While overall data center demand stays strong, this means the Q2 outperformance was mainly driven by hyperscale cloud providers, while enterprise, industrial, and some AI cloud-related demand grew slower than previously expected.

By contrast, edge computing business revenue in Q2 was $7.2 billion, up 27% year-over-year and 13% quarter-over-quarter, beating market expectations of $6.61 billion.

This segment covers RTX, Windows client-side AI, robotics, autonomous driving, etc. Its scale is much smaller than data centers and forms the core of Nvidia’s “physical AI” narrative. While revenue keeps growing, in the short term it is still difficult to change Nvidia’s heavy dependence on data center revenue.

Nvidia expects Q3 revenue of $108 billion, with a +/-2% range, i.e., about $105.8 to $110.2 billion—a year-over-year increase of 85.6% to 93.3%, with the midpoint up 89.5%. Even at the low end, the company will enter the “$100-billion-quarter” club.

This guidance is above consensus estimates. Different institutions put the Q3 consensus at ~$104 to $105 billion; Nvidia’s midpoint of $108 billion means another 3-4% upgrade.

However, this is also a key reason for after-hours stock pressure: the “real hurdle” for Nvidia may be much higher than consensus estimates.

Before the results, the most optimistic LSEG consensus was $112 billion. Against such buy-side expectations, the $108 billion guidance is strong, but not further expanding the imagination space.

Also, Q3 guidance midpoint grows about 12% from Q2, lower than Q2’s 18% sequential growth. As the base rises rapidly, slower growth is not surprising, but for a high-valuation stock, any change in sequential growth margin is magnified by the market.

Q2’s 75.0% gross margin, flat with Q1, is a solid result—especially with revenue near $100 billion and stressed data center supply chains. But Q3 gross margin guidance is 74.0% (plus/minus 50 basis points), down about one point from Q2.

This is not severe deterioration. A ~74% margin remains extremely high for the semiconductor industry globally and shows Nvidia continues to have strong bargaining power. For investors, the question is whether gross margin is near a cyclical high point.

Several factors could impact future margins: the initial ramp of the Vera Rubin platform may bring cost and yield fluctuations; costs for HBM, advanced packaging, and substrates keep rising; and as full-rack system sales grow, product mix may differ from pure GPU sales.

Thus, the market’s sensitivity to Q3's gross margin decline does not reflect worsening profitability, but that Nvidia’s valuation already prices in “high growth + high margins sustained or improved.”

On expenses, the company expects Q3 GAAP operating expenses ~ $9.2 billion, non-GAAP operating expenses ~ $9 billion, both above Q2’s $8.408 billion and $8.232 billion. With Rubin mass production, network product expansion, AI software, and “AI factory” ecosystem push, R&D and operating investments continue to accelerate.

The key product signal in this report is that Vera Rubin is entering full-fledged mass production—meaning Nvidia is shifting market focus from the Blackwell cycle to the Rubin cycle.

According to disclosures, Rubin is not a single chip but an entire AI factory architecture that includes Vera CPU, Rubin GPU, Spectrum-6 networking, BlueField, security, storage, software toolchains, and the DSX platform. Nvidia repeatedly uses the “AI factory” concept to highlight that it is supplying not just chips, but complete systems for building, running, and scaling AI compute infrastructure.

This is important for the valuation narrative. The market previously worried whether Nvidia could smoothly transition to the next-generation platform after the Blackwell demand peak. If Rubin can scale production as management says and quickly deploy among major cloud and AI infrastructure companies, it will help extend the growth cycle.

But platform transitions also bring short-term uncertainty. New platform ramps usually come with changes in supply chain, delivery cadence, customer acceptance, and cost structure. Q3 guidance for lower gross margin may also turn some investor focus to profit margins during Rubin’s early mass production phase.

Nvidia made it clear that Q3 guidance "does not assume any data center computing revenue from China." This is a crucial detail.

On the positive side, the $108 billion guidance does not include any contribution from Chinese data center compute. If future regulatory, export license, or product arrangements improve, the China business could be an upside factor.

From a trading perspective, it also reminds the market that Chinese data center revenue remains uncertain in the short term. For a company with over 90% of revenue from data centers, any regional restriction is going to be factored by investors as a risk discount.

Thus, excluding China revenue makes the guidance “cleaner” and more conservative but also means the market cannot count on a recovery from China driving Q3’s certain growth.

In Q2, Nvidia’s free cash flow was $21.341 billion—still an impressive figure but noticeably down from $48.554 billion in Q1. The main reason isn’t weakening profitability but rising working capital usage.

The cash flow statement shows receivables were up $22.346 billion, inventory up $5.784 billion, and prepayments plus other assets up $5.497 billion. This reflects business expansion, growing client deliveries, and supply chain stocking, but also means profit-to-cash conversion did not keep pace this quarter.

At the end of Q2, cash and cash equivalents were $22.443 billion, with short-term debt at $1 billion and long-term debt at $32.366 billion. In Q2, the company returned about $26 billion to shareholders through share buybacks and cash dividends, with $19.7 billion for buybacks and ~$6 billion for dividends. As of quarter-end, about $99 billion in buyback authorization remains.

Capital returns are not weak, but there were previous calls for Nvidia to emulate Apple and ramp up buybacks even further. The report only disclosed remaining authorization and regular dividends, with no more aggressive capital return commitment beyond current authorization, possibly disappointing some investors.

In the report, Nvidia announced strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR to mobilize over $500 billion in third-party capital for AI infrastructure, subject to finalized agreements.

This represents Nvidia’s attempt to address AI infrastructure bottlenecks: compute demand needs not only GPUs but also power, land, data center shells, networking, cooling, and long-term financing. The company also disclosed cooperation with SB Energy at Ohio’s PORTS-Pike Technology Campus to secure land, power, and shell capacity for hosting Nvidia compute resources.

This is also the most debated part of the market. Investors will ask: how much future AI infrastructure spending will come from customer capital expenditures versus vendor/private equity/credit support? Will Nvidia shoulder potential credit support, residual risks, or long-term obligations?

Nvidia CFO Colette Kress defended the company’s aggressive AI investment strategy in the earnings call. She said Nvidia expects that top frontier AI labs like OpenAI “will become the largest technology companies in history.”

Kress noted these labs face bottlenecks in securing all the compute needed for developing and optimizing models, hence their need for Nvidia's financial support: "We are acutely aware of the size of this support and know some may call it ‘circular financing.’ But we see it differently. The capital we put in will provide excellent equity returns.”

From the balance sheet, Nvidia’s investment assets are growing rapidly.

At the end of Q2, marketable equity securities were $42.783 billion and unlisted securities were $51.157 billion, totaling about $93.9 billion—up substantially from the start of the year. Q2 equity securities net gain was about $7.77 billion, with a six-month cumulative net gain of ~$23.7 billion. These investments contribute to GAAP profits but also make the market focus more on the relationship between ecosystem investments and core revenue.