On Friday, October 16, CoreWeave's shares fell 11.91% in a single session, and the Nasdaq Composite closed down 2.14%. The drop capped a week in which the AI stock selloff spread from chipmakers to the debt-funded companies that build and rent out the data centers those chips power. CNN reported that the index sank more than 2% as investors questioned how much borrowing the AI boom can support.
The total damage is large, even if the exact tally depends on the source. According to a CNBC report relayed by AI Weekly, 20 of the most valuable global AI and semiconductor stocks lost more than $1 trillion in value during the week. The weekly decline for the Nasdaq itself is disputed. Reuters said the index was on track for a loss of more than 4%, while BigGo Finance reported a drop of more than 5%.
Stocks steadied on Monday, October 19, and the Nasdaq climbed as tech shares rebounded, according to TheStreet. Whether that bounce holds is still open. The next real test comes with late-October earnings from Big Tech, when investors will look for signs of how much those companies plan to spend on AI infrastructure.
Why the AI stock selloff hit so hard
The reversal was sharp. The Nasdaq and S&P 500 set record highs about ten days before the October 16 drop, according to CNN, so the fall came quickly after a strong run.
The central worry is financing. Neoclouds such as CoreWeave rent AI computing capacity to labs and big technology firms, and they fund their growth with debt. When Treasury yields rise, that borrowing gets more expensive and harder to refinance. Daniela Hathorn, a senior market analyst at Capital.com, said in AOL's Stock Market Today column that the sell-off reflected "renewed concerns about the financing behind the AI boom."
Chip demand is a separate question. One Killua Capital Substack analysis cites Broadcom's Hock Tan as reportedly saying that capital, not chips, may be AI's biggest constraint. If that view holds, the risk is not that buyers stop ordering processors. The risk is that the companies renting that hardware cannot comfortably pay for it.
Which stocks took the biggest hits in the sell-off
| Stock or index | Move on Oct. 16 | Why it matters |
|---|---|---|
| CoreWeave (CRWV) | Down 11.91% | Debt-funded neocloud that rents GPU capacity |
| Broadcom (AVGO) | Down 4.52% | AI-related chip supplier under renewed pressure |
| Nvidia (NVDA) | Down 1.43% | Chip supplier that partly backs CoreWeave |
| Philadelphia Semiconductor Index | Down 3.83% | Broad read on chip stocks |
| Nasdaq Composite | Down 2.14% | Broad tech benchmark |
| VIX volatility index | Closed at 19.99, up 21.45% | Market fear gauge |
The chip index fell harder than the Nasdaq, which points to semiconductors as the center of the rout. CoreWeave sits at the center of the financing story. Its stock trades as a proxy for both the chip cycle and the debt cycle, and Nvidia's partial backing ties the two together.
Nvidia's own move was modest on the day. A Substack analysis of the selloff estimated that Nvidia lost about $606 billion in market value between Oct. 6 and Oct. 16, and that the total loss across the 12 stocks it tracked reached $1.03 trillion. Nvidia has kept releasing models, including the open Nemotron 3.5 Super VL, a 120-billion-parameter model built for AI agents. Still, the week's selling centered on financing rather than product news.
The pain also reached Asia. South Korea's Kospi posted its worst week since 2008, falling more than 14% on AI worries, according to CNBC. The index also sat about 40% below its June high.
Debt, rates and the cost of building AI data centers
Admiral Markets, citing TIKR, put CoreWeave's net debt at about $46 billion, or roughly 8.2 times trailing EBITDA, in its analysis of the selloff. That is a heavy load for any company, and it matters more when borrowing costs climb. The sources reviewed reported no default, and nothing in the coverage shows that CoreWeave's debt has become a credit event.
Rates are the other half of the story. The Federal Reserve raised its policy rate 25 basis points on Sept. 16, to a range of 3.75% to 4.00%, according to a Neo Reports brief. On Oct. 16, Fed Chair Kevin Warsh said at the IMF and World Bank meetings that the Fed would not pre-judge its October or December decisions. Whether another hike is coming is uncertain, and that uncertainty weighs most on heavy borrowers.
The spending on the other side of the ledger is enormous. Estimates for 2026 across major buyers add up to roughly $720 billion to $745 billion, and Moody's expects hyperscaler capital spending to approach $1 trillion in 2027, according to Admiral Markets. That demand is why the financing question matters, because every dollar of new capacity has to be paid for somehow.
What the CPI report and the OpenAI revenue debate add
Inflation data added pressure. Analysts linked the sell-off partly to a hotter-than-expected US CPI report, which pushed Treasury yields higher and renewed doubts about how AI data centers are financed, according to AOL and Yahoo Finance. A separate Yahoo Finance analysis argued that the sell-off could catch bearish investors off guard.
OpenAI's revenue figure became a flashpoint. Reports cited an annualized revenue figure of about $50 billion, below the $68 billion to $70 billion that had circulated. Reuters also reported that OpenAI itself indicated a figure near $70 billion, so the sources still disagree. The discrepancy matters because AI labs are the customers that neoclouds depend on to justify building capacity.
A related valuation story landed the same week. One market analyst noted that the drop began before a Reuters report that Anthropic's pre-IPO valuation had slipped to $1.5 trillion from $2 trillion. USA 24 has covered the Anthropic valuation talk separately.
AI bubble or blip: how to read the rebound
The sources do not agree on how to measure the week. BigGo described it as the Nasdaq's biggest weekly drop in a year, while a CNBC live-blog headline called it the Nasdaq's worst week since June. Investing.com also examined which tech stocks held up during an earlier AI selloff in early October, so this is not the first AI-driven pullback this month.
The case for a blip is straightforward. Stocks rebounded on Monday, and the major indexes had reached records only days before the drop. The case for a deeper correction rests on financing. If borrowing costs keep rising and neocloud debt gets harder to refinance, companies may slow data-center projects, and that would hit the chip suppliers that depend on those orders.
Who loses is clearer than who wins. Holders of index funds and AI-themed ETFs absorb the paper losses, borrowers face higher costs as yields rise, and chip suppliers depend on spending continuing. Consumers and businesses that buy AI services could also face pricing pressure if AI companies cut capital spending or delay projects. That outcome is an inference, not a confirmed one.
What to do if your portfolio holds AI stocks
The reporting points to a few practical checks. This is general information, not personal financial advice.
- Avoid trading on a single day's move. The Nasdaq fell on Oct. 16 and steadied on Oct. 19.
- Check overlap across your funds. Index funds and AI-themed ETFs often hold the same names, so look for Nvidia, Broadcom and CoreWeave across all your holdings.
- Measure concentration. Add up what you own in semiconductors and AI infrastructure, including through funds, and compare that total with your target allocation.
- Set your rebalancing rules and cash reserves before the next move, not during it.
- Mark the calendar. The Fed meets Oct. 27–28, and Big Tech earnings arrive later in October.
What comes next for AI spending and the Fed
The next two weeks will settle the near-term question. The Fed's Oct. 27–28 meeting will show how policymakers weigh inflation against the cost of financing, and it sets the backdrop for late-October Big Tech reports. Investors should watch two things in those reports: whether capital spending plans hold, and whether companies explain how they fund data-center buildouts. If spending holds and borrowing stays manageable, Monday's rebound may look like a pause. If either breaks, the financing story will move from analyst notes into earnings guidance, and the selloff could have a longer tail.
