AMD Q2 2026: Data Center Revenue Doubles to $6.7B as Gaming Drops 31%

AMD reported Q2 2026 data center revenue of $6.7 billion, more than doubling the $3.2 billion it posted in the same quarter a year ago and climbing from $5.8 billion in Q1 2026. AI infrastructure demand is driving the surge. CEO Lisa Su told investors the data center segment is on track to more than double year-over-year again in 2027. On the other side of the ledger, AMD's gaming business fell 31 percent to $779 million, dragged down by price increases and component shortages that slowed console sales.

What happened

Metric	Figure
Q2 2026 data center revenue	$6.7 billion
Q1 2026 data center revenue	$5.8 billion
Q2 2025 data center revenue (year-ago)	$3.2 billion
Year-over-year data center growth	107%
Q2 2026 gaming revenue	$779 million
Gaming revenue change (year-over-year)	-31%

AMD’s data center business, which sells chips and accelerators (purpose-built processors for AI training and inference workloads) to cloud providers and enterprises, has now more than doubled its revenue in a single year. The $6.7 billion Q2 result is not a one-quarter spike: it follows $5.8 billion in Q1, suggesting sustained demand rather than a lumpy pull-forward.

Speaking on the earnings call Tuesday, CEO Lisa Su projected that the data center segment will again more than double year-over-year in 2027. That is an aggressive forecast that would put annual data center revenue well above $25 billion if the trajectory holds.

Why is AMD’s gaming revenue falling?

The gaming segment reported $779 million for the quarter, down 31 percent from the same period last year. AMD attributed the drop to price increases and component shortages that weighed on sales of the Xbox Series X/S, PlayStation 5, and Valve’s Steam Deck. All three use AMD silicon, so weaker console sell-through flows directly into AMD’s gaming line.

This is not purely an AMD problem. It reflects broader pressure on consumer hardware, where elevated prices are squeezing unit volumes across the industry.

Why it matters

The 107 percent data center growth rate confirms that AI chip demand has not cooled at the scale buyer level. AMD is the most direct public competitor to NVIDIA in AI accelerators, so its results serve as a useful second data point on the health of that market. When both companies are reporting strong AI infrastructure numbers, it signals that enterprise and cloud spending on AI capacity is still accelerating, not plateauing.

For businesses building AI-powered products or evaluating AI infrastructure costs, this level of demand tends to keep GPU compute prices elevated. If you are budgeting for cloud AI services in 2026 or 2027, plan for continued pricing pressure rather than relief. Our coverage of open-weight AI models is worth reading alongside this if you are looking for ways to reduce dependence on expensive cloud inference.

The gaming decline is a reminder that AMD’s consumer and enterprise businesses are moving in opposite directions. For hardware resellers or game developers, soft console demand is a real near-term headwind.

Our take

Lisa Su’s 2027 forecast is bold and will get a lot of attention, but the Q1-to-Q2 sequential growth from $5.8 billion to $6.7 billion is actually the more credible data point right now. It shows the ramp is not stalling between quarters. That matters because AI infrastructure buildouts often front-load orders, and sustained sequential growth is harder to maintain than a single year-over-year comparison.

The gaming drag is real but not surprising. Console cycles are long, and this one has hit a rough patch on supply and pricing. It does not change AMD’s overall story, which is increasingly an AI and data center story.

If your business is planning AI integrations or evaluating chip vendors, AMD’s trajectory confirms this is a multi-year cycle, not a short burst. For practical next steps on putting AI infrastructure to work in your own operations, see what AI integration looks like at the project level.

What to do about it

1. Budget for elevated AI compute costs through at least 2027 based on AMD’s own demand forecast.
2. Evaluate open-weight or on-premise model options if cloud inference costs are a line item you want to control.
3. Watch AMD’s Q3 2026 report for whether sequential data center growth continues past $6.7 billion, which would validate Su’s 2027 projection.
4. If you sell consumer hardware or develop for consoles, factor in continued softness until price and supply conditions improve.

The practical takeaway: AMD’s numbers make a strong case that AI infrastructure spending is a sustained cycle, so build your cost and vendor assumptions accordingly rather than waiting for prices to drop.

Source: The Verge · AI