Photo Serge Nussbaumer © Maverix
By Serge Nussbaumer, Head of Public Solutions and financial markets expert at Maverix
Nvidia once again beat Wall Street’s already lofty expectations. In the second quarter of fiscal 2027, revenue jumped 106% to $96.2 billion, versus a forecast of around $92.2 billion.
Adjusted earnings per share came in at $2.22, also above consensus.
But the outlook matters most. For the current quarter, Nvidia expects revenue of roughly $108 billion, while analysts had been looking for just over $104 billion. Sales of data-center computing products in China are still excluded from these figures.
The stock initially reacted with restraint, but rose sharply during the earnings call and was up about 5% by the close. The market therefore ultimately interpreted management’s comments as confirmation that the AI investment cycle is still not losing momentum.
The data-center business remains the main growth engine. Revenue there increased 117% to $89.0 billion and now accounts for more than 92% of group revenue. For now, there is no sign of any slowdown in investment in AI infrastructure.
In addition, Nvidia has already begun developing its next generation of chips, “Vera Rubin.” Alongside CUDA, networking and software, this fast product cycle is one of the company’s core competitive advantages.
Profitability is equally extraordinary. Gross margin reached 75%, while operating income jumped 124% from a year earlier to $63.7 billion. The combination of triple-digit growth and such high margins remains Nvidia’s key strength.
The most critical issue concerns margins. For the coming quarter, Nvidia is guiding for a gross margin of around 74%. Given the sharp rise in memory and component costs, that figure could even fall to between 71% and 72% in the fourth quarter.
Even so, that level remains exceptional. And there is an important counterpoint. Before the quarterly results were even published, reports said Nvidia had informed customers of price increases of more than 15% on certain AI server systems. The steep rise in memory and other component costs is seen as the main reason. Nvidia also suggested it intends to charge higher prices for its products from the start of the next fiscal year.
The expected margin decline therefore takes on a different meaning. If Nvidia manages to pass a meaningful share of those additional costs on to customers, the margin pressure would be temporary rather than evidence of a structural erosion in its pricing power.
That is precisely where one of the strongest messages in these figures may lie. If Nvidia can impose double-digit price increases on already extremely expensive AI systems without any visible drop in demand, its pricing power remains exceptional.
The long-term competition therefore comes from its own customers. Google, Amazon, Microsoft and Meta are investing billions in their own AI accelerators to reduce their reliance on Nvidia. In inference in particular, specialized chips may ultimately prove a more economical alternative to high-end general-purpose GPUs.
In the near term, however, there is little to suggest Nvidia is losing its market leadership. CUDA, networking, software and hardware together form a formidable technological moat. In addition, the annual product cycle makes it difficult for competitors to close the technology gap.
The key long-term question is therefore not so much whether Nvidia will remain the market leader. It is more likely that the company will succeed in defending that position. What matters more is which prices and margins Nvidia will still be able to command in a more mature and competitive AI market.
Beyond all this operational strength, the balance sheet also deserves attention. Since the end of January, accounts receivable have risen from $38.5 billion to $63.1 billion, and inventories from $21.4 billion to $31.6 billion.
At the same time, GAAP net income of $59.7 billion was matched by operating cash flow of only $24.1 billion. When revenue doubles, a significant increase in working capital is not unusual. Given Nvidia’s financial strength, this is not a warning sign for now. Still, it is something to watch, especially if receivables and inventories continue to grow faster than the business.
In conclusion, these results clearly reinforce the AI investment thesis. A company posting quarterly revenue of nearly $100 billion is still growing by more than 100%, while revenue of $108 billion is already expected for the next quarter. And that excludes any meaningful revenue from data centers in China.
The after-hours share-price move of about 5% is therefore understandable. Expectations were extremely high, yet Nvidia still managed to beat them.
Margin trends remain the main operational weak spot in the near term. But the recently announced price increases show that Nvidia is trying to pass higher costs on to customers. Whether that effort succeeds will be an important test of the company’s true pricing power.
The investment thesis is therefore evolving. The decisive question is no longer whether demand for AI infrastructure is increasing. It is rather how much of the value created Nvidia will be able to capture over the long term.
At present, there is little to suggest that Nvidia’s dominant position is under immediate threat. The combination of growth, technological lead and pricing power remains exceptional. But the bigger Nvidia gets, and the higher prices rise, the more customers will try to reduce their dependence.
That is the central conflict in Nvidia’s story. Its enormous pricing power is today one of the company’s key assets – and at the same time the strongest economic incentive for its customers to develop alternatives tomorrow.
Retrouvez l’ensemble de nos articles Strategic Case
