The AI trade had its worst week in months. Between Thursday, October 29 and Friday, October 30, Alphabet suffered the worst single day in its history, Intel lost more than a quarter of its value, and the Nasdaq Composite fell 2.86%, its biggest one-day drop since June. The question everyone is now asking is simple: has the AI bubble finally burst?
The short answer: not yet, at least not by the numbers. But the market has clearly changed how it judges AI spending, and that shift matters for every company building with or selling AI.
Key takeaways
- Alphabet fell about 13% on October 29, erasing roughly $500 billion in market value after its search revenue missed estimates and free cash flow turned negative.
- On October 30, Intel dropped 28.6% on weak guidance, and the Nasdaq closed at 25,437, about 7.8% below its October 6 record.
- Chip stocks were hit hardest: AMD fell 10.9%, Dell 10.8% and Nvidia 4.2% on Friday.
- The common thread is not that AI demand has vanished, but that investors are punishing companies whose AI spending is pushing cash flow negative.
- A drop of under 10% from a record is a correction in progress, not a confirmed crash. Whether it becomes one depends on earnings and spending news in the coming weeks.
What happened
Thursday: Alphabet’s record drop
The trigger was Alphabet’s third-quarter report. Overall revenue grew 25% to $127.6 billion and Google Cloud revenue doubled to $30.7 billion, but search, the company’s core profit engine, grew 12% to $63.4 billion, short of the roughly $64.9 billion analysts expected. Google executive Philipp Schindler said AI Mode queries, which answer questions conversationally, are eating into traditional searches and leading to lower ad monetisation, according to the Irish Times.
Spending made it worse. Alphabet’s capital expenditure hit $51.3 billion in the quarter, and the company said full-year capex could reach $205 billion. Free cash flow came in at about negative $9.1 billion, only the second negative quarter since its 2004 IPO. Investors reacted by sending the stock down roughly 13%, the worst day on record for the company and a loss of around $500 billion in market value.
Friday: the selloff spreads
Late Thursday, Amazon and Intel reported. Amazon’s revenue and earnings per share both came in slightly below forecasts, even though AWS growth accelerated to 39%. It kept its 2026 capex forecast near $220 billion while free cash flow was negative, according to Charles Schwab’s market update. Intel guided fourth-quarter revenue to $15–16 billion, well below expectations, despite strong growth in its data centre and AI business.
By Friday’s close, according to Reuters data:
- Nasdaq Composite: down 2.86% to 25,437
- S&P 500: down 1.82% to 7,580.07
- Dow Jones: down 1.09% to 51,770.31
- Intel: down 28.6%
- AMD: down 10.9%; Dell: down 10.8%; Nvidia: down 4.2%
- Semiconductor index: down 7.2% on the day and more than 15% for October
Reddit also fell about 29% after its CEO said Google search referrals had dropped sharply in October, a reminder that AI-driven search is reshaping traffic across the web, a trend we explored in our piece on zero-click search. Crypto stocks fell too, but that was linked to a reported hack at Binance rather than AI.
So, has the AI bubble burst?
Calling this a burst bubble goes further than the evidence allows, for three reasons.
First, the size of the move. The Nasdaq sits about 7.8% below its record from early October. Markets usually call a 10% drop a correction and 20% a bear market. Painful as this week was, the broad index has not yet reached correction territory, let alone crash levels.
Second, AI demand is still growing. Google Cloud revenue doubled. AWS growth sped up. Intel’s data centre and AI revenue rose sharply even as the stock collapsed. The selloff is not about customers abandoning AI; it is about whether the returns arrive fast enough to justify the spending.
Third, the market is discriminating. Microsoft, which reported strong positive free cash flow alongside heavy capex, held up far better than Alphabet on Thursday. Schwab’s analysts summed up the pattern: investors appear to reward AI spending funded from cash flow and punish spending that drives free cash flow negative. A true bubble collapse tends to drag everything down indiscriminately.
That said, the warning signs are real. Big Tech is committing hundreds of billions of dollars a year to data centres, chips and energy, financed increasingly with debt. Alphabet alone has future commitments for data centres and related infrastructure of more than $1 trillion, according to the Irish Times. If revenue from AI products does not catch up, those commitments become a burden. As Wealthspire Advisors’ Oliver Pursche told Reuters, the market’s question now is whether there will be a positive return on investment, and how quickly.
Why it matters
This week marks a change in the story Wall Street tells about AI. For most of the past two years, bigger AI budgets were rewarded almost automatically. Now spending has to show up in cash flow. That puts pressure on every link in the chain: hyperscalers may slow or stretch out capex, chipmakers and server makers face tougher questions about future orders, and AI startups may find investors more demanding about revenue and margins. We looked at how far AI funding had run in our analysis of Q1 2026 AI investment; a correction in public markets usually reaches private valuations with a lag.
There is also a less obvious twist in Alphabet’s numbers. Google’s own AI product is cannibalising its most profitable business. If AI answers reduce the number of ad-supported clicks, the economics of search, and of the publishers who depend on it, change permanently.
What it means for developers and businesses
- AI tools are not going away. The cloud providers’ AI businesses are growing fast, and the infrastructure already built will be used. Expect continued model releases and price competition.
- Pricing could tighten. If providers are pushed to show returns, generous free tiers and heavily subsidised API pricing may become less common. Budget for that.
- Startups should focus on revenue. In a more cautious market, a clear path to paying customers will matter more than a big AI narrative.
- Avoid single-vendor lock-in. Building with portable tools and multiple model providers reduces risk if any one provider changes pricing or strategy.
What to watch next
- Earnings from Apple and Palantir on Monday, AMD on Tuesday, and Arm and Qualcomm on Wednesday.
- The US midterm elections on Tuesday, November 3, and the October jobs report on Friday, November 6.
- Whether any hyperscaler trims its capex plans, which would be a much stronger bubble signal than one bad week of share prices.
- Whether the Nasdaq falls more than 10% from its record, officially entering correction territory.
- Nvidia’s next earnings report, the clearest read on whether AI chip demand is holding up.
The bottom line: the AI boom has not popped, but the free pass is over. From here, AI spending has to pay for itself.
This article is for information only and is not financial advice.