Nvidia Earnings Ahead with China, AI Spend in Focus

August 25, 2025
Nvidia, the biggest AI-chip firm, reports Wednesday. Watchlist items include the pathway toward resuming H20 chip sales in China, revenue guidance, and Blackwell growth.

Nvidia (NVDA) approaches Wednesday afternoon's earnings report in much better shape than three months ago when it suffered a severe backlash from tariffs and restrictions on chip exports to China. Nvidia fell briefly below $90 per share in early April and was just starting to recover in May when the company last reported. 

Since then, the stock soared to new all-time highs near $180—more than double its April trough—as the company deftly worked through its trade issues, and the Trump administration backed off the most restrictive tariffs against China. 

Though Nvidia is just one stock, its earnings carry lots of weight in the S&P 500®  index (SPX), meaning strength or weakness in its Q2 results for fiscal 2026 could have a broad impact. 

"Going into this earnings report, no company (since 1981 when records began) has ever accounted for a larger percentage of the S&P 500's market capitalization, which means this is no longer just a quarterly earnings release but a macro market event that warrants attention, even if you aren't directly invested in Nvidia shares," said Alex Coffey, senior trading and derivatives strategist at Schwab. 

Nvidia's market capitalization recently reached $4.4 trillion, nearly equaling the market cap of the entire S&P health care sector and representing more than 8% of the S&P 500's total value. 

A year-to-date chart of Nvidia (NVDA) shares through August 19 shows the roller-coaster ride this year. Shares are now about double the level of their April low, when the stock was haunted by tariff fears and concerns about possible AI competition from China.

Nvidia stock has done so well over the last few weeks that it may take a blowout earnings report to get investors much more excited. Shares sported a forward price-to-earnings (P/E) multiple above 42 by mid-August, according to research firm CFRA, and Nvidia's market capitalization has more than doubled in just over a year. The SPX has a forward P/E of just above 22, by comparison. 

A high P/E ratio isn't necessarily a hindrance if a company can back it up with increasingly strong earnings. And so far, Nvidia has. But potential challenges include a margin slowdown over the last year, accompanied by concerns that the law of large numbers might make the sort of stratospheric growth of 2022–2024 harder to achieve. Sales more than doubled in fiscal 2024 and are expected to again in fiscal 2025. They're then expected to rise more slowly at 52% and 22% in fiscal 2026 and fiscal 2027, respectively, according to CFRA research. 

Sliding gross margins played into the headwinds that hurt shares earlier this year. They moderated to the low 70's on a percentage basis from the high 70's seen in early 2024. The question is whether Nvidia's continued rollout of its more advanced Blackwell chip can bring margins back up in this quarter and future ones. Any guidance from Nvidia on that might be something to watch. 

Other things to check when Nvidia reports are any updates on its pathway to sell H20 chips in China now that it's approved to do so by the Trump administration and general revenue guidance now that the trade barrier has eased. 

"I am looking forward to seeing updates to the revenue guidance now that there is additional clarity and a pathway to resuming sales of H20 chips and perhaps other models to China after the massive adjustment to projections last quarter," Coffey said. 

New chip generates hope for profit gains

Blackwell, introduced over the last year, is a graphics processing unit (GPU) that Nvidia promotes as improving AI inference. AI inference refers to pretrained AI models being deployed to generate new data. Inference can be used for anything from training driverless cars to obey traffic rules to helping farmers improve their crop production. As Blackwell sales increase and push aside sales of older chips to possibly grow Nvidia's already strong market share, margins might benefit. 

Pulling back for a wider view, Nvidia's earnings offer a chance for investors to glean whether the heavy spending each quarter by so-called "hyperscaler" cloud providers like Microsoft (MSFT), Meta Platforms (META), and Alphabet (GOOGL) is getting transferred into Nvidia's vault and what trends the company sees moving ahead. The hyperscaler spending has been unprecedented and provides solid evidence of AI's usefulness for these big cloud providers as they ramp up their data center networks, but most observers think it will slow at some point. When is uncertain, but Nvidia may have insight. This year, Meta alone intends to spend up to $72 billion on AI infrastructure, and Alphabet recently raised its 2025 capital spending forecast by $10 billion to $85 billion. 

Even when cloud spending diminishes, Nvidia won't necessarily be left behind, considering recent developments in both the corporate and government spheres. When Nvidia shares hit record levels above $180 recently, it partly reflected Tesla's (TSLA) decision to disband its Dojo supercomputing team and associated in-house training chips and rely more on hardware from Nvidia and Advanced Micro Devices (AMD), Barron's reported. "It doesn't make sense for Tesla to divide its resources and scale two quite different AI chip designs," Tesla CEO Elon Musk said in a post at the time. 

That came around the same time that OpenAI CEO Sam Altman named Nvidia as one of its key partners as it launched GPT-5, the latest model powering ChatGPT, Barron's noted. 

China sales in focus as barriers ease

Positive news also flashed for Nvidia last month when President Trump announced the U.S. government would once again allow sales of Nvidia's H20 chips—a less-advanced product—to China. Previously, U.S. security restrictions influenced Nvidia's decision to give up on those sales, which at the time CEO Jensen Huang called "deeply painful" and said could ultimately cost Nvidia $15 billion. The company took a steep charge in its fiscal first quarter related to wiping those sales off the books, but now it might be able to sell at least some of these chips in China. 

How many isn't clear, however, as China discouraged local firms from buying the products, Bloomberg reported, particularly for government-related purposes. Beijing wants Chinese firms to buy domestic chips instead, though U.S. AI chips are coveted for certain applications. Nvidia, which earlier this month agreed to a controversial plan to share 15% of its revenue from Chinese sales of H20 with the U.S. government, might bring more clarity to these issues when it hosts its earnings call Wednesday. It's likely to be asked about both by analysts. 

Drilling down into quarterly numbers, Nvidia reported fiscal first-quarter data-center revenue growth of 10%, up 73% from a year ago. That far surpassed 14% annual data-center growth recently reported by rival AMD, a small number that caused worries about AMD's ability to keep up with Nvidia. However, if Nvidia sees data-center revenue growth slow in the second fiscal quarter, it might raise concerns about an industry slowdown. AMD told analysts that its AI chip revenue was hurt by U.S. export restrictions on China and the rollout of its next-generation AI chips, Reuters reported. 

Nvidia guided for second-quarter revenue of $45 billion, plus or minus 2%, and gross margins of 72% on a non-Generally Accepted Accounting Principles (GAAP) basis. Both numbers will be closely watched because Nvidia has a track record of beating guidance and analysts' expectations. As of mid-August, analysts expected second-quarter revenues of $45.7 billion, up from just over $44 billion in the first fiscal quarter and up about 52% year over year. 

Guidance closely watched as investors look for strength

The level of Nvidia's beat on quarterly revenue guidance—should it do so, as analysts expect—could help determine the stock's path Thursday when the market opens. The company's $44 billion revenue in the first quarter was about $700 million, above average expectations on Wall Street. 

The consensus earnings per share (EPS) estimate for the second quarter was recently $1, up from $0.81 in the first quarter and $0.68 a year ago. The first-quarter EPS reflected the loss of H20 sales in China that resulted in a $4.5 billion charge associated with excess inventory and purchase obligations. Excluding the H20 charges and related tax impact, EPS in the first quarter were $0.96, above the average analyst estimate of $0.93. 

In an effort to maintain aggressive growth objectives, Nvidia is working to both cultivate demand and launch new initiatives. Earlier this year, it announced strategic partnership with HUMAIN—the new full AI value chain subsidiary of Saudi Arabia's Public Investment Fund—to build AI infrastructure in the Gulf country. Factories in Saudi Arabia with a projected capacity of up to 500 megawatts will be powered by several hundred thousand of Nvidia's most advanced GPUs over the next five years. 

Nvidia emphasized how deals like these reduce its dependence on hyperscalers. Many analysts following the industry say the heavy spending can't continue forever, and AI firms like Nvidia will eventually feel the impact of lower demand. It's unclear when that will happen, but sovereign AI arguably provides another revenue stream. 

In addition, Nvidia announced that it's opening its AI server platform to rival chipmakers, including Qualcomm (QCOM). This would allow companies to launch custom central processing units (CPUs) that can link Nvidia's chips to power data centers. 

The AI business remains Nvidia's centerpiece, but it experienced a recent revival in its smaller gaming business. First-quarter gaming revenue rose 48% to $3.8 billion, up 42% from a year ago. Automotive revenue remained a minor handicap in the first quarter, falling 1% from a quarter earlier. 

Some things to listen for in Nvidia's call Wednesday afternoon include: 

  • What adjective will CEO Huang use to describe AI-chip demand? Last time he opted for "very strong" but has used more colorful language in the past.
  • What is the status of Nvidia's Rubin chip? This came under scrutiny in mid-August when a Fubon Research analyst wrote that he thinks it's "very likely" that Rubin will be delayed, saying Nvidia is redesigning the chip to better match a rival offering from AMD. The market expects mass production of Rubin to begin by the third quarter of 2025, Barron's said.
  • How will Nvidia describe its guidance for sales to China in light of its recent agreement with the Trump administration to share revenue from those sales, and will that revenue sharing have any impact on margins? 

All expressions of opinion are subject to change without notice in reaction to shifting market, economic or political conditions. Data contained herein from third party providers is obtained from what are considered reliable sources. However, its accuracy, completeness or reliability cannot be guaranteed. 

This material is intended for general informational and educational purposes only. This should not be considered an individualized recommendation or personalized investment advice. The investment strategies mentioned may not be suitable for everyone. Each investor needs to review an investment strategy for his or her own particular situation before making any investment decisions. 

All corporate/ issuer names and market data shown are for illustrative purposes only and are not a recommendation, offer to sell, or a solicitation of an offer to buy any security. 

Investing involves risk, including loss of principal. 

Past performance is no guarantee of future results. 

0825-1ACW