What Sparked the Slide
Two big forces are colliding: pricier money and pricey AI build-outs. This week, a basket of 20 leading AI and chip companies has seen over $1 trillion erased as rate hikes, questions around how to pay for AI infrastructure, and mounting competitive pressure soured the mood.
A hotter U.S. inflation print on Wednesday stoked worries the Fed could push rates higher again, setting up a tough Thursday. On Wednesday, the Nasdaq Composite fell 1.88% to 26,692.87, the S&P 500 slipped 1.11% to 7,718.94, and the Dow dipped 0.09%.
Financing is front and center. On Monday, the chiefs of Broadcom and Intel cautioned that the cost and availability of funding could be a serious obstacle for the AI boom. Broadcom's Hock Tan said it might prove a bigger challenge to scaling AI than chip or power shortages. Dan Ives of Yorkville Ives noted that with U.S. Treasury yields topping 5%, the price of capital is rising, but he argued that demand for AI compute remains strong.
There's also a fresh worry in memory. Investors are weighing the risk of a global oversupply as soon as 2029 as manufacturers crank up capacity for AI. Over the weekend, China's ChangXin Memory Technologies said it achieved a new architecture that boosts memory density by storing more data in a smaller footprint, underlining intensifying competition. On X, Elon Musk added that China would be making "a LOT of RAM in 2028."
Where the Damage Hit
AI-heavy names fell across the board Thursday. In the U.S., Nvidia slipped roughly 2%, and both AMD and Intel were lower by about 4%. Dell Technologies slid 8%. Micron was down roughly 5% earlier, then turned positive around noon ET.
The "neoclouds" also retreated. CoreWeave fell 3%, Nebius declined 5%, and IREN was off by over 8% but later recouped part of the drop following remarks by Nvidia CEO Jensen Huang. Speaking in Stockholm, Sweden, Huang stated there was no necessity for Nvidia to step in as a backstop for CoreWeave, and added, "Their cost of capital is actually decreasing." A live quote box showed CoreWeave at 67.84, up 3.12%, at 1:53 PM EDT.
The selloff deepened in Asia and Europe. In South Korea, SK Hynix fell by more than 10% and Samsung Electronics declined 8.73%. LG Innotek dropped 8.95%, Seoul Semiconductor lost 9.28%, and Samsung SDI slipped 6.26%.
When innovation meets uncertainty, staying grounded matters, so try the free Market Briefs newsletter for clear daily context
Japan's tech giants were also hit: SoftBank Group tumbled 7.9%, Advantest fell almost 14%, Tokyo Electron declined 6.45%, and Renesas Electronics slid 8%. Taiwan's TSMC dropped 1.81%. In Europe, ASML was last down 5.2%, while Infineon Technologies and ASM International fell 8.2% and 8.6%, respectively.
Rates, Funding, and the AI Build-Out
Investors are growing worried that more expensive borrowing could squeeze the massive capital spending required for AI infrastructure, pressuring the very firms that have ridden the AI investment boom.
Dan Ives put it this way: "Higher yields raise the cost of financing these builds, and investors are right to focus on balance sheets and contract structure, but rates are a financing question, not a demand question." He characterized the pullback as a buying opportunity and said the AI infrastructure cycle is still in its early innings, even with funding costs climbing.
There's a caveat. "What would change our view is a demand signal, not a rate signal," he said, flagging three warning lights to watch: potential reductions in hyperscaler capital spending, contracts being scrapped, or declines in GPU rental rates.
What It Means for Your Portfolio
This week is a reminder that when money gets pricier, long-horizon projects feel it first, and AI is capital hungry. The other swing factor is memory: talk of a possible glut by 2029, a CXMT density leap, and that "a LOT of RAM in 2028" comment all point to supply risks later in the cycle. If inventory outruns demand, some parts of the stack could lose pricing power.
Right now, the market is trading the cost of money more than the appetite for compute. Keep an eye on the demand signals Ives highlighted - hyperscaler capex plans, the cadence of new contracts, and GPU rental pricing - to gauge how durable this AI build-out really is for your own mix of holdings.
Markets move fast and headlines blur together, so get the free Market Briefs newsletter to understand stocks, crypto, and housing
