AMD just posted a record quarter. Nvidia’s data center business is still 10 times bigger.

AMD delivered the kind of quarter that would make most chipmakers jealous. Revenue rose 45% to a record $13.4 billion in the three months to September, the company said on Tuesday, ahead of the $13.08 billion analysts expected. Adjusted earnings came in at $2.09 a share against forecasts of $1.97, and AMD expects revenue of about $17 billion in the fourth quarter, plus or minus 3%.

The shares, which closed up 3.7% at $483.90 before the report, traded about 8% higher shortly after it, according to Yahoo Finance.

Yet set against Nvidia, the numbers show how far AMD still has to go in the race to sell chips for artificial intelligence.

Record growth, from a much smaller base

AMD’s data center segment, which sells both its EPYC server processors and its Instinct AI accelerators, grew 97% to a record $8.5 billion. Chief executive Lisa Su said growth would “accelerate further” in the fourth quarter as AMD scales up Helios, its full-rack AI system, and its next generation of server chips.

Nvidia’s data center business made $89.0 billion in its most recent quarter, the three months to late July. That is more than ten times AMD’s figure, and Nvidia’s quarters end two months earlier, so the comparison flatters AMD if anything. Nvidia has guided to total revenue of about $108 billion for its current quarter.

Data center revenue in each company’s latest reported quarter, billions of dollars

Show the numbers

	Data center revenue
AMD (Jul–Sep 2026)	$8.5bn
Nvidia (May–Jul 2026)	$89bn

Source: AMD and Nvidia earnings releases

Growth rates do not close the gap either. Nvidia’s data center revenue grew 117% from a year earlier, faster than AMD’s 97%. In dollars, Nvidia added roughly $48 billion of quarterly data center sales in a year. AMD added about $4.2 billion. Even while doubling its data center sales, AMD is falling further behind in absolute terms.

Where the gap shows

• Profitability: Nvidia’s gross margin was 75% last quarter. AMD’s adjusted gross margin was 56%, and it expects about 55% next quarter. Nvidia can charge more because customers have fewer alternatives.
• Scale of profit: Nvidia earned $59.7 billion in net income in one quarter, more than four times AMD’s total revenue.
• Market share: Silicon Analysts, drawing on Bloomberg Intelligence and IDC data, estimates Nvidia will take about 75% of AI GPU revenue in 2026 and AMD around 7% to 10%. AMD’s data center figure also includes server CPUs, so its AI accelerator sales alone are smaller still.
• Software: Nvidia’s CUDA platform has nearly two decades of tools and millions of developers behind it. AMD’s ROCm has improved quickly, especially for running models, but still trails for large-scale training.

Why AMD still matters

None of this makes AMD’s quarter a weak one. The company’s largest customers want a second supplier so they are not dependent on one chipmaker for prices and supply. Anthropic, Meta, Microsoft, OpenAI and Oracle are already running workloads on the first Helios systems built around AMD’s new Instinct MI450 chips, and the guidance for the fourth quarter implies growth of roughly 66% from a year earlier.

There were weaker spots. Gaming revenue fell 40% to $773 million, and PC chip sales of $2.9 billion missed the $3.06 billion analysts expected, as higher memory prices hurt demand. On a standard accounting basis, profit was also cut by a $1.1 billion charge linked to warrants AMD has issued to a customer.

What it means

AMD is doing what it needs to do: winning big customers, launching rack-scale systems and growing quickly. But the AI chip market is not yet a two-horse race. For now, Nvidia is the market and AMD is the alternative. The next test comes when Nvidia reports its next quarter in the coming weeks, which will show whether AMD is gaining ground or just growing alongside a leader that is growing faster.