
Looking at the artificial intelligence boom through Nvidia’s results
When it comes to the artificial intelligence (AI) boom, few events command as much attention as Nvidia’s quarterly earnings calls. With so much riding on the boom enduring, including Nvidia’s $US4.15 trillion market cap, analysts and traders understandably dissect every nuance – from the precise wording of forward guidance to the subtle inflections in CEO Jensen Huang’s voice. They’re looking for clues about whether excitement or caution lies ahead or is presently justified.
Nvidia’s earnings reports are arguably the single most influential quarterly disclosure for the entire market. And not just the Nasdaq – the broader market too.
In an era where artificial intelligence is being hyped as driving unprecedented economic transformation (we’ve heard that before – Metaverse anyone?), Nvidia sits at the epicentre, the picks and shovels supplier in the race to discover Artificial General Intelligence (AGI), powering everything from generative AI models to massive data centres.
If the company’s results exceed expectations, investors can breathe a collective sigh of relief, knowing they can dream big – at least for another quarter. Conversely, with valuations stretched, a hint of weakness in Nvidia’s results or outlook will ripple through the S&P 500 and Nasdaq.
Pre-earnings jitters
Heading into the second quarter earnings report, which was released on 27 August, 2025, investors were concerned about three issues.
First were the ongoing, if not escalating, U.S.-China trade tensions, which have cast a long shadow over Nvidia’s operations. Indeed, the company recently struck a deal with the U.S. government, agreeing to hand over 15 per cent of its revenue from China in exchange for export allowances on certain chips. More recent reports, however, indicate Nvidia has paused production for China amid Beijing’s concerns about product security and specifications. With tensions increasing following the Chinese Communist Party’s (CCP) military parade and exhibition, questions remain about how Nvidia will navigate any disruptions in one of its key growth geographies.
Second, investors were eager for insights into future demand from the hyperscalers – Microsoft, Amazon, Alphabet, and Meta, who are pouring billions into AI infrastructure. The current tech bull market owes much of its existence to the massive spending spree on AI training and deployment by these companies. Concerns of a slowdown from a shift in capital priorities or from economic headwinds had emerged, prompting concerns about whether hyperscaler demand would be sustained.
Finally, there was the perennial question of valuation and momentum. Nvidia’s stock has already surged more than 30 per cent year-to-date on optimism about its persistent AI dominance. But with investors now trained to expect the company to routinely smash expectations, a mere “beat” won’t suffice.
The numbers
When the dust settled on the July quarter (Nvidia’s fiscal second quarter), the headline figures painted a picture of robust, if not explosive, performance, fueled by the ongoing AI boom but tempered by external pressures.
Revenue climbed 56% per cent year-over-year to $46.7 billion, roughly in line with analyst estimates. Net income jumped 59 per cent to $26.4 billion. These gains reflect the AI ecosystem’s continued expansion, with surging demand from AI developers.
The company’s all-important data centre segment, which now accounts for 89 per cent of Nvidia’s total sales, grew 56 per cent annually to $41.1 billion. This category encompasses the company’s powerhouse chips used for training and refining AI models. However, sequentially from the prior quarter, data centre revenue edged up only 5 per cent, a figure that may indicate slowing momentum.
Looking ahead, Nvidia projected third-quarter sales of $54 billion – slightly above the Wall Street consensus but falling short of loftier forecasts by some analysts who had pencilled in $60 billion. Bloomberg noted the discrepancy; that while the guidance met average expectations, it underwhelmed those anticipating another blockbuster.
The key factor dragging on the outlook? The ongoing international trade frictions, particularly with China, which we have already noted. Nvidia disclosed it isn’t anticipating any sales of its H20 chips –specially designed to navigate U.S. export restrictions – to Chinese customers in the current quarter.
Despite the earlier agreement to remit 15 per cent of H20 sales from China to the U.S. for export permissions, Beijing has reportedly advised its domestic firms to steer clear, citing security concerns and dissatisfaction with the chips’ downgraded capabilities. As a result, Nvidia halted production for this market and excluded any resumption of H20 sales from its forward guidance.
In the July quarter, the company recorded no new H20 chip sales to China, though it did fulfil $180 million in previously ordered units there, plus another $650 million to an unrestricted customer outside the region. The absence of these sales contributed to a $4 billion shortfall in data centre revenue, marking the second straight quarterly miss for the segment (although the previous one occurred before analysts fully adjusted for China-related losses).
CFO Collette Kress indicated that if geopolitical hurdles clear, Nvidia could potentially add $2 billion to $5 billion in third-quarter sales from H20 resumption, with even more upside from fresh orders.
Market reaction
Nvidia’s shares had notched a fresh all-time high leading into the report, but the earnings release produced a muted response, stemming largely from the “underwhelming” revenue projection, fanning fears that AI chip demand might be approaching a plateau after quarters of dizzying growth. The narrow miss on data centre expectations amplified the jitters.
Jensen Huang’s vision: This time is different
CEO Jensen Huang offered a predictably bullish narrative, painting his company not just as a chipmaker but as a comprehensive AI infrastructure provider. Once the pioneer of the graphics processing unit (GPU) for gaming, the company has expanded into servers, networking, and full data centre solutions over the past decade, all aimed at maximising efficiency.
“We’re really an AI infrastructure company,” Huang stressed. He noted the largest AI players could invest $3 trillion to $4 trillion over the next five years based on their current capex trajectories, with Nvidia potentially capturing up to 70 per cent of the market.
Huang also noted a 17 per cent sequential increase in sales of the new Blackwell line of graphic processing unit (GPUs_ – Nvidia’s most advanced yet – describing demand as “extraordinary.” Early adopters include Disney, Hitachi, Hyundai Motor, and SAP for Blackwell-equipped servers.
Despite the China setbacks, Nvidia’s core business currently remains insulated by AI fervour. But it is also sensitive to changes in that fervour. Nevertheless, comments by OpenAI, Microsoft, Amazon, Alphabet, and Meta suggest there appears to be no slowing in appetite for Nvidia’s chips.
Shareholders might also cheer the announcement of an additional $60 billion stock buyback, up from the $24.3 billion bought back in the fiscal year’s first half (with $14.7 billion still authorised at quarter’s end).
Huang capped the earnings call with optimism: “This year is a record-breaking year. I expect next year to be a record-breaking year as well.”
The cracks in AI’s halo
Nvidia’s results are a litmus test for the AI boom. The company’s dominance has made it a proxy for AI’s transformative potential, often dubbed “the fourth industrial revolution”, reshaping every sector and even humanity itself.
On the plus side, the sustained 56 per cent growth in data centre revenue confirms AI spending by hyperscalers is not slowing. If Huang’s $3-4 trillion forecast materialises, it would reflect a multi-year capex supercycle, benefiting not only Nvidia but also suppliers, cloud providers, and AI software firms. This could buoy broader indices, supporting economic growth through innovation and productivity gains.
Yet, rarely has technology capable of changing the course of human history also generated wild profits for investors. More often than not, a boom becomes a bubble, and the subsequent losses render only consumers of the technology the winners.
For Nvidia, the softer guidance and China uncertainties introduce caution. The sequential slowdown in data centre growth, even if partly attributable to one-off factors, might signal that hyperscalers are digesting their massive investments, are reprioritising or facing budget constraints amid higher interest rates. A perceived plateau in AI demand could trigger a re-evaluation of tech stocks, potentially leading to sector-wide pullbacks.
Geopolitically, the U.S.-China saga also highlights vulnerabilities in global supply chains. If tensions escalate, as they appear set to under evolving U.S. policies, Nvidia’s exposure could drag on performance, with knock-on effects for competitors like Advanced Micro Devices and the semiconductor ecosystem.
And then there’s sentiment; investors accustomed to Nvidia’s outperformance may demand even greater beats to maintain enthusiasm, raising the bar for future quarters.
The law of large numbers, of course, virtually guarantees a slowdown at some point.
The Polen Capital Global Growth Fund own shares in Nvidia, Microsoft and Amazon. This article was prepared 04 September with the information we have today, and our view may change. It does not constitute formal advice or professional investment advice. If you wish to trade these companies you should seek financial advice.