Twenty of the most valuable AI and semiconductor stocks in the world lost more than $1 trillion in combined value so far this week, according to CNBC's reporting on Thursday, October 15, 2026. South Korea's SK Hynix fell more than 10% in one session. Japan's Advantest dropped almost 14%.
The more interesting detail in the reporting is that the stated cause was the cost of money, not a drop in AI usage. This post breaks down what moved, why analysts say it happened, and how to read a market move like this without panic or hype. It is not investment advice.
Quick summary: CNBC reports 20 major AI and chip stocks lost over $1 trillion so far that week, as of October 15, 2026, after hotter-than-expected US inflation data revived rate-hike fears and Treasury yields above 5% raised concerns about funding AI data centers. Analyst Dan Ives called it a financing question, not a demand question. Exact percentage moves differ slightly between outlets.
What moved on Thursday
CNBC lists Thursday moves in Asia: SK Hynix down more than 10%, Samsung Electronics down 8.73%, LG Innotek down 8.95%, SoftBank Group down 7.9%, Advantest almost 14% lower, Tokyo Electron down 6.45%, Renesas down 8%, and TSMC down 1.81%. In Europe, Infineon fell 5% and ASM International 7%. (Source: CNBC)
Gotrade's closing figures differ a little, which is common when outlets report at different times: SK Hynix down 11.06%, Samsung down 9.48%, SoftBank down 9.33%, and the Kospi down 7.57% to 5,554.23 after dropping as much as 9% intraday and triggering a 20-minute trading halt, which Gotrade says was the tenth circuit breaker on the index this year. It puts the Nikkei 225 down 4.90% to 64,425.90. (Source: Gotrade)
On Wall Street the previous session, both outlets agree the Nasdaq Composite fell 1.88% to 26,692.87 and the S&P 500 fell 1.11% to 7,718.94. CNBC's Thursday US moves were mixed: Nvidia down 1.84%, AMD down 1.5%, Dell down 6.8%, while CoreWeave rose 5.94% after early losses and Micron gained 1.87%.
Why it happened
CNBC says tech led losses after hotter-than-expected US inflation data raised the prospect of further Federal Reserve rate hikes, with US Treasury yields above 5%. Gotrade puts September US consumer price inflation at 3.7% year on year versus 3.4% in August, and the 10-year yield at 5.11% early Thursday, after 5.35% earlier in the month, which Gotrade, citing AP, calls the highest since 2002. The logic is that AI infrastructure is financed with borrowed money, so higher rates raise the cost of data centers and equipment.
CNBC adds three other threads. Investors worry capacity expansion could create a memory chip glut as early as 2029. China's ChangXin Memory Technologies said it had a breakthrough in a new architecture that boosts memory density. And CNBC says the CEOs of Broadcom and Intel warned on Monday that financing could become a major obstacle to AI expansion, with Broadcom's Hock Tan saying it could be a bigger challenge than chip or electricity shortages.
| Market | CNBC | Gotrade | Note |
|---|---|---|---|
| SK Hynix | More than -10% | -11.06% | Memory exposure |
| Samsung Electronics | -8.73% | -9.48% | Timing of figures differs |
| SoftBank Group | -7.9% | -9.33% | AI investor |
| Nasdaq Composite | -1.88% to 26,692.87 | -1.88% to 26,692.87 | Prior US session |
| S&P 500 | -1.11% to 7,718.94 | -1.11% to 7,718.94 | Prior US session |
| CoreWeave | +5.94% Thursday | Not covered | Rebound after early losses |
What analysts and Nvidia's CEO said
Dan Ives of Yorkville Ives told CNBC the selloff reflects concern about the cost of financing infrastructure, while underlying AI compute demand stays strong: "rates are a financing question, not a demand question." He called the pullback a buying opportunity, and said what would change his view is a demand signal: cuts to hyperscaler capital spending plans, canceled contracts or falling GPU rental prices. (Source: CNBC)
Neocloud stocks rebounded after Nvidia CEO Jensen Huang told Bloomberg in Stockholm that Nvidia did not need to backstop CoreWeave, adding "Their cost of capital is actually decreasing." Morningstar's Lorraine Tan told Gotrade that any risk to perceived AI spending growth could slow chip and equipment demand, hitting tech-heavy markets like Korea and Taiwan. Remember that these are opinions, and a single volatile week can reverse.
What it means if you just use AI tools
- Nothing in the reporting says chatbot usage fell. The stated worry is how AI buildouts get funded.
- Expensive money can slow data-center expansion, which is worth watching for capacity limits and pricing over time, but no outlet reports a price change from this event.
- The three demand signals Ives named (capex cuts, canceled contracts, falling GPU rental prices) are the ones that would matter more than rate moves.
- Market swings do not tell you which AI model is best for your task.
Do not tie your workflow to one vendor's pricing or availability. Keep a second and third model you already know how to prompt, so a price change or outage never stops your work.
01How much did AI and chip stocks lose?
CNBC reports that 20 of the most valuable global AI and semiconductor stocks lost more than $1 trillion combined so far that week, as of October 15, 2026.
02What caused the selloff?
CNBC cites hotter-than-expected US inflation, Treasury yields above 5% and worries about the cost of financing AI infrastructure, plus memory oversupply and competition concerns.
03Did Nvidia fall?
CNBC reports Nvidia down 1.84% on Thursday. Gotrade reports a 3.16% fall to $221.30 in the prior US session.
04Is AI demand shrinking?
Analyst Dan Ives says no, calling it a financing question, not a demand question. That is one analyst's view.
05Why do the numbers differ between outlets?
Outlets report at different times and for different sessions, so closing and intraday figures vary slightly.
If you rely on AI tools and don't want to depend on one vendor, LumiChats offers 40+ premium models side by side under one subscription, which makes it easy to keep a backup model you already know how to prompt.
