demand continues to be strong, but the market wanted proof
In the relentless race to power the world's artificial intelligence ambitions, Advanced Micro Devices found itself humbled this week — not by failure, but by the gap between expectation and reality. The company's first-quarter forecast of $9.8 billion, while technically strong, fell short of the most optimistic Wall Street projections, triggering a 16 percent stock collapse that stripped away months of investor confidence. It is a familiar human story: a company doing well, yet measured against a dream it has not yet fully inhabited. AMD's moment may still be coming, but markets, like impatient prophets, punish those who make the future wait.
- A single forward-looking number — $9.8 billion — was enough to erase months of stock gains in a single trading session, exposing how fragile AI-era valuations truly are.
- Despite a genuinely strong quarter — data center revenue up 39%, earnings per share well ahead of consensus — investors fixated on what AMD couldn't promise rather than what it had delivered.
- Nvidia's continued dominance looms over every AMD earnings call like an unspoken verdict, and the market's 16% punishment reflects anxiety that AMD may never close that gap fast enough.
- China adds a volatile undercurrent: AMD's $390 million in older chip sales to Chinese customers is expected to crater to $100 million this quarter, as the company waits on export licenses it does not yet hold.
- CEO Lisa Su's assurances of tens of billions in AI revenue by 2027 and the coming MI325 processor launch offer a credible roadmap — but Wall Street is no longer willing to pay today for a tomorrow that keeps receding.
Advanced Micro Devices saw its stock fall 16 percent in a single session this week — its worst single-day drop since October 2018 — after issuing a first-quarter sales forecast of roughly $9.8 billion. The number beat the average analyst estimate, but it fell short of the higher projections circulating on Wall Street, and it carried an uncomfortable subtext: AMD's artificial intelligence business is not accelerating as fast as the market had hoped.
The underlying quarterly results were genuinely solid. Data center revenue surged 39 percent to $5.38 billion, PC sales rose 34 percent, and earnings per share came in well ahead of consensus. But when executives turned to the outlook, the enthusiasm drained from the room. The guidance painted a picture of a company still running second to Nvidia in the AI infrastructure race, still waiting for the transformative revenue wave that investors had priced in.
CEO Lisa Su struck an optimistic tone, pointing to major partnerships with OpenAI, Oracle, and the Department of Energy as proof of genuine demand for AMD's MI-series accelerators. She projected tens of billions in AI revenue by 2027 and expressed confidence in the upcoming MI325 processor launch in the second half of this year. The market, having heard similar promises before, wanted more than confidence — it wanted a timeline.
China complicated the picture further. AMD moved $390 million in older MI308 chips to Chinese buyers last quarter, a useful revenue buffer — but one that is expected to shrink to around $100 million this quarter as the company awaits export licenses it has not yet received. Negotiations with Washington and potential Beijing customers continue, but no resolution is guaranteed.
With Intel already having disappointed investors on supply constraints, AMD had seemed positioned to capture defecting customers. Its own cautious guidance suggested it wasn't seizing that opening as forcefully as hoped. By Wednesday afternoon, the stock's 13 percent year-to-date gain had largely vanished, leaving investors to wonder whether AMD's AI reckoning would arrive on schedule — or keep them waiting a little longer.
Advanced Micro Devices watched its stock plummet 16 percent in a single day this week—the steepest drop since October 2018—after telling investors that first-quarter sales would land around $9.8 billion. The number itself wasn't catastrophic. It beat the average analyst estimate of $9.39 billion. But it fell short of the higher projections some on Wall Street had circulated, and more importantly, it signaled something the market had begun to fear: AMD's artificial intelligence business wasn't accelerating as fast as the hype had promised.
The chipmaker's fourth quarter had looked strong. Data center revenue climbed 39 percent to $5.38 billion, crushing the $4.97 billion analysts expected. Personal computer sales rose 34 percent to $3.1 billion. The company posted $1.53 in profit per share, well ahead of the $1.32 consensus. But when executives opened their mouths about what came next, the enthusiasm evaporated. The forward guidance suggested a company still playing second fiddle to Nvidia in the race to dominate AI infrastructure, still waiting for the big payoff that venture capitalists and tech investors had been banking on.
CEO Lisa Su remained characteristically upbeat during the earnings call, insisting that demand for AI chips remained robust and that the company would hit tens of billions in AI revenue by 2027. She dismissed concerns about supply chain bottlenecks and said AMD could scale production to meet whatever orders arrived. But the market had heard similar promises before. What investors wanted was proof—concrete evidence that AMD's newer MI325 processor, due in the second half of this year, would actually win significant business away from Nvidia's dominant position.
One complication clouding the picture was China. AMD had shipped $390 million worth of older MI308 chips to Chinese customers in the fourth quarter, a lifeline that helped pad the top line but also revealed the company's desperation to move inventory in a market where it still lacked permission to sell its latest technology. The company expects that number to drop to roughly $100 million this quarter, a decline that underscores how dependent AMD has become on export licenses it doesn't yet have. The Trump administration has signaled a willingness to relax restrictions on chip sales to China, but the Department of Commerce has moved slowly in issuing the necessary permits. AMD said it continues negotiating with Washington and potential customers in Beijing, but there's no guarantee those talks will yield results.
The company has signed major deals with OpenAI and Oracle, and the Department of Energy has also become a customer. These partnerships suggest real appetite for AMD's MI series accelerators, which compete directly with Nvidia's offerings in data centers. But appetite and actual revenue are different things. Wall Street wants to know when the tens of billions in potential sales will actually materialize, and AMD hasn't given them a timeline that satisfies.
Intel, AMD's traditional rival in processors and graphics chips, had already disappointed investors with its own forecast last month, citing supply constraints that prevented it from meeting demand. The market interpreted that as good news for AMD—evidence that the company was stealing share. But AMD's own guidance suggested the company wasn't capitalizing on that opportunity as aggressively as some had hoped. The stock had climbed 13 percent this year through Tuesday's close. By Wednesday afternoon, much of that gain had evaporated, and investors were left wondering whether AMD's AI moment would ever actually arrive.
Notable Quotes
There's no question that demand continues to be strong. And so we're working with our supply chain partners to increase supply as well.— CEO Lisa Su
The Hearth Conversation Another angle on the story
Why did the stock fall so hard if AMD actually beat the average analyst estimate?
Because Wall Street had built a narrative around AMD as the AI chip challenger, and the guidance suggested that narrative was moving slower than expected. Some analysts had projected $10 billion or higher. When you miss the high end of expectations, especially in a sector as hyped as AI, the market punishes you.
But the data center business grew 39 percent. Isn't that exactly what you'd want to see?
It is, and it shows real momentum. But growth rates matter less than what comes next. Investors were betting on an acceleration—that AI spending would drive exponential growth. A 39 percent quarter is strong, but if the guidance suggests that growth is moderating, the story changes.
What's the deal with the China sales? That seemed like a bright spot.
It's complicated. Yes, AMD shipped $390 million of older chips to China, which helped revenue. But those are becoming obsolete. The company expects that to drop to $100 million next quarter. It's a one-time boost that masks a deeper problem: AMD can't sell its newest, most profitable chips to China because of export restrictions. They're waiting for licenses that may or may not come.
So the real story is that AMD is stuck waiting for permission from Washington?
Partly. But it's also that Nvidia is still winning the race. AMD has newer chips coming in the second half of the year, and they've signed deals with OpenAI and Oracle. The question is whether that's enough to close the gap. The market clearly thinks it isn't—not yet, anyway.
Lisa Su said demand is strong. Why didn't investors believe her?
Because executives always say demand is strong. What matters is whether that demand translates into orders and revenue. The guidance suggested it might take longer than Wall Street wanted to wait. In a market this hot, patience is a luxury.