Nasdaq 100 Falls 10.4% in a Week, Its Worst Stretch Since the Pandemic
The US tech sector is having its worst week since the pandemic. The Nasdaq 100 fell 5.6% on Friday and 10.4% over the week. The index is now 20% below its October peak, which places it in a bear market, according to the Financial Times. Investors are worried about doubts over multibillion-dollar AI buildouts, reports that the SEC is examining Nvidia's accounting, and high interest rates. Teams building AI products should plan for a more cautious compute budget.
AI-processed from The Decoder; edited by Hamidun News
The AI market had spent weeks waiting for investments in data centers and chips to pay off quickly. This week, that expectation cracked. According to the Financial Times, as relayed by The Decoder, the Nasdaq 100 fell 5.6% on Friday and lost 10.4% over the full week. It was the worst weekly performance since the 2020 crisis. The index is already 20% below its October high, which is the formal threshold for a bear market.
For those building AI products, these are not abstract market figures. Investor sentiment shapes compute prices, the terms of funding rounds, and how generously labs subsidize their APIs.
What Fell
The biggest blow hit Nvidia, whose shares lost 14% on Friday after reports that the SEC is examining the company's accounting. The source provides no details about the review, so it is too early to judge how serious it is. But the market's reaction shows how strongly the valuation of the entire industry depends on a single accelerator supplier.
The US semiconductor index lost 10.7% in a single day. Memory and chip makers were in the red, including Sandisk, Micron, and Broadcom. Investors are also debating whether the high valuations of these companies are justified, given that the memory and chip markets are cyclical. In plain terms: demand for hardware surges and then collapses, and paying for it as if growth were permanent is risky.
Anthropic, whose shares had traded for only its second day on the exchange, fell 9.1% after its debut. For a newly public company, a drop like this in its first days is painful: it has no trading history yet to smooth out volatility.
Why Everything Happened at Once
The publication describes the situation as a perfect storm of several factors:
- doubts about multibillion-dollar AI buildouts, with no clear timeline for when they will generate returns;
- the scandal surrounding Nvidia and questionable deals with China;
- political risks following the Democratic victory in the midterm elections;
- expensive oil due to the war with Iran, which keeps inflation high and prevents interest rate cuts.
The last point matters especially for capital-intensive projects. Data centers are built largely with borrowed money, and the higher the rates, the more expensive each new gigawatt of capacity becomes. When doubts about returns and the cost of money rise at the same time, investors start demanding evidence rather than promises.
What Developers and Founders Should Do
First, don't draw conclusions from a single week. The source is a short note based on Financial Times reporting, and it offers no forecast of how long the decline will last. A bear market does not mean that AI products have stopped working or that customers have stopped paying.
Second, it is worth checking your product's dependence on a single model provider and a single type of compute. If major players start cutting costs, the first things to go will be generous pricing tiers, free quotas, and startup discounts. Calculate in advance how your margins would change if the price per token rose, and whether you have a fallback on another model or on open weights.
Third, founders planning to raise money should prepare for tougher questions about revenue and unit economics. A market skeptical of projects worth hundreds of billions is unlikely to be softer on startups with polished decks and no paying customers.
Finally, watch two things: how the SEC review of Nvidia ends, and how Anthropic's shares perform in the coming weeks. These two stories will show whether a weekly panic turns into a lasting shift in sentiment or whether the market recovers part of its losses. For now, the practical takeaway is simple: build a buffer into your compute budget and don't assume that cheap access to capacity will stay that way forever.
- Nov 22, 2026Android work profile closes real-IP leak that Happ and v2rayNG don't catch
- Nov 22, 2026Anthropic Goes Public on Nasdaq: Shares Rise from $83 to $111, but Can Safety Hold Up?
- Nov 22, 2026Databricks Is in Talks for a Round at a $225 Billion Valuation, Nearly 20% Above August
- Nov 21, 2026Anthropic and Claude Claim Proof of the Riemann Hypothesis, Stanford Responds Without AI
Key Questions Answered
What Fell?
The biggest blow hit Nvidia, whose shares lost 14% on Friday after reports that the SEC is examining the company's accounting. The source provides no details about the review, so it is too early to judge how serious it is. But the market's reaction shows how strongly the valuation of the entire industry depends on a single accelerator supplier. The US semiconductor index lost 10.7% in a single day. Memory and chip makers were in the red, including Sandisk, Micron, and Broadcom.
Why Everything Happened at Once?
The publication describes the situation as a perfect storm of several factors: doubts about multibillion-dollar AI buildouts, with no clear timeline for when they will generate returns; the scandal surrounding Nvidia and questionable deals with China; political risks following the Democratic victory in the midterm elections; expensive oil due to the war with Iran, which keeps inflation high and prevents interest rate cuts. The last point matters especially for capital-intensive projects.
What Developers and Founders Should Do?
First, don't draw conclusions from a single week. The source is a short note based on Financial Times reporting, and it offers no forecast of how long the decline will last. A bear market does not mean that AI products have stopped working or that customers have stopped paying. Second, it is worth checking your product's dependence on a single model provider and a single type of compute. If major players start cutting costs, the first things to go will be generous pricing tiers, free quotas, and startup discounts.
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.