The Market Just Sent Big Tech Its AI Bill: Over $1 Trillion Gone in a Day, and Apple Back on Top. What Is Real, and What Is Fear

By Thorsten Meyer | November 1, 2026 | ThorstenMeyerAI.com

Buying for a business?Offer from Amazon

Get business pricing on tech for your team

• Business-only prices and quantity discounts
• Tax-exempt purchasing
• Multiple users, one account, clear invoices

Create a free Business account →Free to register

As an affiliate, we earn on qualifying purchases.Hide

For three years, Wall Street rewarded almost every AI announcement. Spend more on data centers, and your stock went up. Last week, that deal broke. On Thursday, 29 October, Alphabet had the worst trading day in its history. On Friday, tech stocks lost more than $1 trillion in market value in a single session. And by Friday’s close, the most valuable company in the world was no longer the company that sells the chips for AI. It was Apple, a company that has largely stayed out of the AI spending race.

The trigger was not a scandal or a crash in demand. It was a bill. Four companies, Amazon, Alphabet, Microsoft and Meta, spent about $200 billion on buildings, servers and equipment in just three months, roughly twice as much as a year earlier. Alphabet and Amazon have already said they will spend more next year. Investors finally started to ask the question every household asks after a big purchase: when does this pay for itself?

I went through the companies’ own earnings releases and calls to separate the part of this story that is real from the part that is fear. Here is what I found, what it means for companies and for users, and where my reading could be wrong. Nothing here is investment advice.

Key takeaways

• The selloff is real and dated. Alphabet’s Class A shares fell 12.6% on 29 October, wiping out about $500 billion, according to Fortune and NBC News. On 30 October, tech stocks lost more than $1 trillion in value, the Nasdaq Composite fell 2.8% and the S&P 500 fell 1.8%, NBC reported.
• Apple passed Nvidia at Friday’s close. Bloomberg, as published by The Straits Times, put Apple at $5.03 trillion and Nvidia at $4.93 trillion at the close on 30 October.
• The bill is real, too, and it comes from the companies themselves. In the third quarter, Amazon spent $77.3 billion on property and equipment, Alphabet $51.4 billion, Microsoft $39.4 billion and Meta $33.9 billion. Each figure is roughly double the year before.
• Three of the four burned cash. Alphabet’s free cash flow was minus $9.1 billion in the quarter, Meta’s minus $2.1 billion, and Amazon’s minus $42.4 billion over twelve months. Microsoft stayed positive.
• The demand is also real. Google Cloud revenue grew 103%, Microsoft’s Azure 48% and AWS 39%. What nobody can prove yet is whether the returns arrive fast enough. That gap is where the fear lives.
• The bigger question is the one The Atlantic raises: is AI money crowding out other investment? US data show data center construction up 73% in a year while factory construction fell about 20%. That fits the argument, but it does not prove it.
• For companies and users, the bill will not stay with Big Tech. Expect more ads inside AI answers, firmer prices and terms for AI services, and tighter competition for construction, power and capital.

Products Worth Considering

LG UltraGear 32GX850A-B 32-inch 4K UHD Glossy OLED Gaming Monitor

AmazonView Latest Price

Samsung 49-Inch Odyssey G93SC QD-OLED Curved Gaming Monitor

AmazonView Latest Price

Samsung 32-inch Odyssey OLED G8 (G80SH) 4K QD-OLED Gaming Monitor

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

Three days that changed the mood

The week started with optimism. Then the earnings came in, one after another.

• Wednesday, 28 October, after the close. Alphabet and Microsoft reported. Microsoft’s cloud grew strongly, and its cash flow stayed positive. Alphabet’s revenue grew 25% to $127.6 billion, but its cash flow turned negative.
• Thursday, 29 October. Alphabet’s Class A shares fell 12.6%, the biggest one-day drop since the company began trading in 2004, Fortune reported. NBC News called the loss of about $500 billion the second-worst one-day wipeout of value in market history. After the close, Amazon and Intel reported.
• Friday, 30 October. Tech stocks erased more than $1 trillion in value, NBC News reported, and Fortune said Nasdaq 100 companies alone lost more than $1 trillion. Intel fell 29% after a weak forecast. The S&P 500’s tech sector dropped more than 4.2%. Amazon recovered from a much deeper fall after hours and closed down 1.4%.
• Friday’s close. Apple’s shares fell 1.2% and Nvidia’s 4.2%. That left Apple at $5.03 trillion and Nvidia at $4.93 trillion, according to Bloomberg, as published by The Straits Times. It was the first time Apple had overtaken Nvidia since late July.

The month as a whole looked just as rough. The Nasdaq 100 fell 6.7% in October, its worst month since March 2025, according to Bloomberg. Money ran to safety. Reuters, as published by The Straits Times, reported that the 10-year US Treasury yield fell to 4.91% in Asian trading on Friday, down about 0.37 percentage points over October.

One detail is easy to miss. On Thursday, while Alphabet crashed, about 400 stocks in the S&P 500 actually rose, and the equal-weight version of the index, where every company counts the same, gained 1.2%, according to Bloomberg, as published by The National. The money did not leave the stock market. It moved away from the AI spenders. The Financial Times described the same shift on Saturday under the headline “Investors trim growth bets over fears of US economy’s AI exposure.” I could not read the full FT article, so I rely on its headline and on the market data above.

The bill, from the companies’ own books

Market prices are opinions. Earnings releases are facts the companies have to stand behind. So here is the bill as the four companies themselves reported it for the third quarter of 2026, which ended on 30 September:

• Amazon: $77.3 billion in purchases of property and equipment, against $35.1 billion a year earlier. Amazon still expects about $220 billion for 2026. CEO Andy Jassy told analysts the company expects “to spend meaningfully more in 2027,” as quoted by NBC News.
• Alphabet: $51.4 billion, against $24.0 billion. Its 2026 guidance of $195 to $205 billion is unchanged, and the company told analysts it continues “to expect our CapEx to increase significantly in 2027.” About 65% of its spending on technical infrastructure went to servers and networking, which typically last about six years. The rest went to the buildings, which last much longer.
• Microsoft: $39.4 billion in cash additions to property and equipment, against $19.4 billion. This figure does not include equipment Microsoft leases.
• Meta: $33.9 billion, or $34.9 billion if you add payments on finance leases, against $18.8 billion. Meta expects $135 to $145 billion for 2026.

Together, that is about $202 billion in one quarter, roughly twice the $97 billion of a year earlier. The definitions differ slightly from company to company, so treat the sum as an order of magnitude, not a precise number. I explained how this money is financed, with cash, bonds, leases and new shares, in Who pays for the AI data center boom?.

What changed the mood is not the spending alone. It is the cash. Free cash flow is the money a company has left after paying for its operations and its investments. It is what pays for dividends, buybacks and debt.

• Alphabet: minus $9.1 billion in the quarter. The company said a stockpile of chips for its TPU system sales and a European Commission fine also weighed on it.
• Meta: minus $2.1 billion, against plus $10.6 billion a year earlier.
• Amazon: minus $42.4 billion over the last twelve months, against plus $14.8 billion a year before.
• Microsoft: positive. Its cash from operations of $58.0 billion covered its $39.4 billion of capital spending, by my calculation from its release. Even so, Microsoft sold $25 billion of bonds last week, its first US investment-grade bond sale in more than nine years, Reuters reported.

What is real, and what is fear

The fairest way to read last week is to split it into three piles.

Real: the spending. The numbers above come from the companies. So do the plans to spend even more in 2027. Banks expect the total to keep rising. Reuters reported that Morgan Stanley sees AI infrastructure spending above $1 trillion next year, and that J.P. Morgan expects $900 billion in 2026 alone. Those are forecasts, not facts, but they point the same way.

Also real: the demand. Google Cloud’s revenue more than doubled to $30.7 billion. Microsoft’s Azure grew 48%. AWS grew 39% to $46.0 billion, and Amazon says it expects growth above 40% in both 2027 and 2028. That last part is Amazon’s own projection. These businesses rent out exactly the computing power the money is buying. If AI were a mirage, these lines would not look like this.

Fear, or more precisely, the unknown: the payback time. Here is the honest problem. A server bought today has to earn its money back within about six years, before it is outdated. Nobody can show yet that it will. Alphabet even said one of its fastest-growing products earns less for now: its finance chief told analysts that strong growth in AI Mode queries “had an impact because of the lower monetization we were seeing in AI Mode.” Google is now scaling ads inside AI Overviews and AI Mode to fix that.

Chris Weston, an analyst at the broker Pepperstone, put the mood into one sentence, as quoted by Reuters: “The market isn’t necessarily questioning whether the AI build-out can be financed. It’s increasingly questioning whether shareholders will earn an adequate return on the enormous capital being deployed.”

That is why Microsoft, which earned more than it spent, was treated very differently from Alphabet on the same night. I compared the two in Same night, opposite verdicts. The market did not punish AI. It punished AI spending that is not yet covered by cash.

Why Apple ended up on top

Apple’s rise to the top is the clearest signal of what investors want right now. Apple has stayed out of the data center race. Bloomberg describes it as having “kept a lid on capital expenditures,” and its shares were up 27% in 2026 at Friday’s close, the best performance among the seven largest US tech companies, again according to Bloomberg.

Apple did not win because it sold more AI. It won because it spent less on it. Its new CEO, John Ternus, told Wired, as reported by TechCrunch on 31 October: “We don’t want to ship AI for AI’s sake.” In a week like this one, investors heard that as a feature.

A word of caution. Market value rankings can flip back within days. Apple fell 1.2% on Friday. It only took the top spot because Nvidia fell more. This is a statement about the mood of one week, not a verdict on either company.

The bigger question: what does the economy give up?

On 31 October, The Atlantic published an essay by Annie Lowrey with a provocative title: “What the Economy Is Sacrificing to AI.” Her argument is simple. Every dollar spent on a data center is a dollar not spent somewhere else. Economists call this crowding out. AI money could be lifting an otherwise weak economy, or it could be pulling capital, workers and power away from everything else. She asks which one it is.

Her example is Dallas. According to The Atlantic, about 125 data centers run in the metro area, 19 more are under construction and 53 are planned. At the same time, the number of new houses that builders started fell 12% in a year, and the number of apartments under construction fell 43% over three years. Those Dallas figures are The Atlantic’s. I could not check them against an official source, so I report them as hers.

What I could check is the national picture. The US Census Bureau’s construction spending report, published on 1 October, shows the split clearly for August 2026, at an annual rate:

• Data center construction: $85.0 billion, up 73% from $49.1 billion a year earlier.
• Factory construction: $168.2 billion, down about 20% from $209.8 billion.
• All construction: $2,203 billion, down 1.7%.

So, money is flowing into data centers while the rest of construction shrinks. That fits the crowding-out argument. It does not prove it. Factory construction had already been falling since late 2024, after a wave of new plants, and high interest rates weigh on housing on their own. My reading: the honest answer is that AI is probably doing both: adding activity in some places and competing for the same builders, electricians, power lines and loans in others. I looked at how much of US growth now depends on AI investment in Computers and software drove a third of US growth last quarter.

The September construction report is due on 2 November. It will show whether the split continued.

What this means for companies

• Expect your AI suppliers to collect on their investment. Companies that spend $200 billion a quarter will look for revenue everywhere: higher list prices, fewer discounts, tighter usage limits and more paid tiers. Budget for that now instead of assuming today’s prices hold.
• Check the financial health of critical suppliers. A cloud or AI provider that is burning cash is not about to disappear. But it may change terms, merge products or end services that do not pay. Ask about price protection and exit clauses in long contracts.
• Use the investor test inside your own company. Investors now ask whether AI spending earns its money back. Ask the same of your own AI projects: which ones save time or money you can measure, and which ones only look good in a slide.
• If you build anything physical, plan for competition for the same resources. Data centers compete with offices, factories and housing for contractors, electricians, grid connections and financing. Book capacity early and expect longer lead times.
• If you are an advertiser, watch Google’s move to put ads inside AI answers. Your search campaigns will increasingly run in formats you have not tested yet.

What this means for users

• More ads in AI answers. Google says it is scaling ads within AI Overviews and AI Mode, including in international markets. Learn to spot sponsored results inside AI answers, just as you learned to spot them on a search page.
• Free tiers may get tighter. When companies need AI to pay for itself, generous free plans are usually the first thing to change. Keep copies of important chats and files.
• Your devices may cost more. Intel’s CEO pointed to “the unprecedented memory crunch and the resulting softness in PC demand” on its earnings call, as quoted by NBC News. Memory is scarce partly because AI data centers buy so much of it, and that can show up in the price of your next phone or laptop.
• You may already be invested. Pension funds and index funds hold these companies, and a handful of them make up a large share of the main US indexes. That is not advice, just something worth knowing when you read that $1 trillion disappeared in a day.

Limits of this analysis

• Stock moves and market values are as reported by NBC News, Fortune, Reuters and Bloomberg for 29 and 30 October 2026. Different outlets give slightly different figures, for example 2.8% or 2.9% for the Nasdaq on Friday. I use the most commonly reported value and name the source.
• The capital spending figures come from the companies, but the definitions differ: Microsoft’s excludes leases, Meta’s guidance includes finance-lease payments, Amazon’s is before proceeds and incentives. The four-company sum is my own addition.
• I could read only the headlines of the two Financial Times articles and the opening of The Atlantic’s essay. Their arguments are reported as theirs.
• Bank forecasts and company projections are forecasts, not facts.
• A few days of trading do not show whether there is an AI bubble. Prices can recover quickly or fall further.
• Nothing here is investment advice. I make no forecast for any stock.

What to watch next

• The Census Bureau’s September construction report on 2 November.
• Apple’s next earnings report, which Reuters said was due the week after the selloff, and Nvidia’s, expected in mid-November.
• The 2027 spending guidance from Alphabet and Meta, which both promised details with their fourth-quarter results.
• Whether ads inside AI answers start to close Alphabet’s monetization gap.

Sources

Alphabet, Q3 2026 earnings release (28 October 2026) and Q3 2026 earnings call (28 October 2026).

Amazon, Amazon.com Announces Third Quarter Results (29 October 2026).

Microsoft, FY27 Q1 earnings press release (28 October 2026).

Meta Platforms, Q3 2026 results (21 October 2026).

US Census Bureau, Monthly Construction Spending, August 2026 (1 October 2026), and private construction time series.

NBC News, Tech stocks wipe out $1 trillion in value as concerns deepen about the AI boom (30 October 2026).

Fortune, Wall Street wipes $1 trillion from the stock market as tech’s $828 billion AI spending spree sparks market fears (31 October 2026).

BBC News, US tech stocks see sharp falls as AI spending nears $1tn (30 October 2026).

Bloomberg, Apple Overtakes Nvidia to Become World’s Most Valuable Company (30 October 2026), as published by The Straits Times (31 October 2026); also reported by Notebookcheck (31 October 2026).

Bloomberg, as published by The National (29 October 2026).

Reuters, as published by The Straits Times (30 October 2026) and by CNA (30 October 2026).

Financial Times, US tech stocks tumble as earnings stoke AI spending fears (30 October 2026) and Investors trim growth bets over fears of US economy’s AI exposure (31 October 2026), subscriber access, headlines only.

The Atlantic, Annie Lowrey, What the Economy Is Sacrificing to AI (31 October 2026).

TechCrunch, Apple CEO says the company doesn’t ‘want to ship AI for AI’s sake’ (31 October 2026).

Media reports are marked as reported. Company statements and projections are quoted as their claims. Nothing here is investment advice.