Nvidia continues to crush it

Nvidia reported revenue of $96.2 billion for the quarter ended 26 July, up 106% on a year ago and up 18% on the previous quarter.

Data centre revenue rose 117% year on year to $89 billion and now makes up 92% of total sales, while net income climbed 126% to $59.7 billion against the $51.2 billion analysts had expected.

Adjusted earnings came in at $2.22 per share versus the $2.10 analysts had forecast, and revenue beat the $92.17 billion consensus.The company guided to $108 billion of revenue in the current quarter, plus or minus 2%, ahead of the $104.2 billion Wall Street had pencilled in.

Chief Financial Officer Colette Kress told analysts on the call that Nvidia expects revenue to grow around 70% in fiscal 2028, well above the 44% analysts had been modelling.

Kress said customer forecasts point to growth doubling next year, and that the guidance still reflects supply constraints rather than any softening in demand.Kress put cloud industry backlog at more than $2 trillion, with capital expenditure by the top five hyperscalers expected to reach nearly $800 billion in 2026 and $1.3 trillion the year after.

“AI has reached its inflection point,” said Jensen Huang, Founder and Chief Executive of Nvidia. He said the technology is doing useful work, that its tokens are productive and profitable, and that demand is accelerating.

Gross margin, the share of sales left after the cost of building the chips, landed at 75% in the quarter, and management guided to 74% for the current quarter and a floor of 71% to 72% in the fourth.

Nvidia blamed soaring memory costs that have run past its earlier expectations and are set to climb further next year.

“We want to be direct about this, rather than let it linger as an open question,” said Kress. She said the memory scarcity is driven in large part by the AI buildout itself.

Kress said the memory shortage will act as a brake on growth at least through fiscal 2028, which runs from February 2027, as Nvidia’s demand for high-bandwidth memory spills into standard server DRAM and NAND flash.

Supply commitments have swollen to $279 billion, tied largely to Vera Rubin memory. Nvidia announced an expanded AWS partnership covering a further 2 million GPUs and confirmed that Vera Rubin production shipments have started, with the platform expected to account for around 20% of data centre revenue in the current quarter.

The company shipped less than 1% of total data centre revenue in Hopper 200 products to Chinese customers during the quarter under US government licences, and its forward outlook assumes no China data centre compute revenue at all because of geopolitical uncertainty.

Worries eased

Ben Barringer, Head of Technology Research at Quilter Cheviot, said investors came into the results with a list of worries covering competition, the durability of AI spending, memory costs, margins, financing arrangements, open-source models and China, and that management worked through each of them in turn.

Barringer said Nvidia pointed to AWS as a customer to make its case on competition, arguing its ecosystem still confers a significant advantage even as hyperscalers build their own silicon.

On financing, he said the company stressed that it is building financing platforms rather than lending to customers, and sought to dispel the idea that it is taking on credit risk.

“Overall, this was a strong quarter, but more importantly a strong call,” said Barringer. He said China remains the bigger unresolved question, with little contribution from the region in the current numbers.

The stock was down about 0.5% after the initial release before turning higher in the after-hours session, and closed the extended session up around 4% on the fiscal 2028 forecast.

Now read: More funding coming for UK video game and creative tech businesses in 2027