Published: July 29, 2026 | Reading Time: ~12 minutes | Channel: techminute
Sometime around 4:15 PM Eastern on July 23, the unspoken deal between Silicon Valley and Wall Street evaporated. Alphabet had just raised its capital spending guidance to $205 billion. Cloud revenue was up 82%. By any reasonable measure, Google's AI business was screaming. And yet the stock tanked 7% — the worst day in over a year.
Within hours, $797 billion had been erased from the Magnificent Seven. It was the biggest single-day bloodletting since the tariff panic of April 2025.
Today, Microsoft and Meta report earnings. Tomorrow, Apple and Amazon. Four of the seven companies that have carried the S&P 500 for three years are about to walk into a market that has stopped applauding their spending and started running the math.
The AI capex reckoning isn't coming. It's here.
Let's be clear about what happened with Google: it wasn't a bad quarter. It was an existential one.
Alphabet posted an 82% surge in cloud-computing revenue — the kind of number that, in any other era, would have sent the stock soaring. But investors had already moved on. What they fixated on instead was a single line item: free cash flow, which turned negative for the first time since Google's IPO in 2004. Let that sink in. A company that generates money by simply existing — that monetizes every search, every video, every map query — burned through more cash than it produced.
The culprit? Capital expenditure. Google raised its 2026 capex ceiling to $205 billion. Not a typo. Two hundred and five billion dollars, mostly aimed at AI data centers, custom TPUs, and the infrastructure to keep Gemini competitive.
"People are really focused on capex, obsessed with it," Jason Lemire, chief investment officer at Bold Wealth Partners, told Fortune. "It used to be the more the better, but now it is the less the better."
The market heard him loud and clear.
Let's look at what the "AI capex problem" actually looks like on a spreadsheet, because the scale is genuinely hard to internalize.
| Company | 2026 Capex Estimate | 2027 Capex Estimate | YTD Stock Performance |
|---|---|---|---|
| Alphabet | $175B – $205B | — | Post-earnings rout |
| Microsoft | ~$190B | — | -21% |
| Meta | $125B – $145B | 14 GW AI infra by 2027 | -9.8% |
| Amazon | — | — | ~Flat |
| Combined (4 cos) | ~$724B | ~$950B | — |
| Apple | ~$13B (FY2025) | — | +23% |
The combined 2026 capex of Alphabet, Microsoft, Amazon, and Meta is projected at roughly $724 billion, according to Bloomberg analyst consensus. In 2027, that number climbs toward $950 billion. For context: that's more than the GDP of Switzerland, and nearly double what the US federal government spends annually on defense R&D.
And Apple? Apple spent $12.7 billion on capex in all of fiscal 2025. The company is up 15% in July alone — its best month in exactly three years — and 23% in 2026. The market's message could not be clearer: we're rewarding the company that isn't writing blank checks for GPUs.
There's something almost poetic about what's happening in the chip sector right now.
The Philadelphia Semiconductor Index (SOX) was up 101% through the first half of 2026. It has since shed 17% in July alone — on track for its worst month since June 2022, when the entire market was still reeling from the inflation crisis. Volatility over the past 100 days is at its highest since the pandemic panic of 2020. The SOX has seen 17 daily moves of 5% or more this year, matching the most since the 2008 financial crisis.
Meanwhile, the S&P 500 has had zero 5% days. The broader market is calm. The chip sector is a hurricane.
This makes a kind of brutal sense. Nvidia, AMD, and Micron are the direct recipients of that $724 billion fire hose. Their valuations depend entirely on the premise that the hyperscalers will keep spending — and keep spending more. If Microsoft and Meta come out tonight and signal even a hint of restraint? The chip trade unwinds fast.
"There is going to be an AI winter at some point," Lemire warned. "When you look at how exceptional margins are — especially in memory — well, it is impossible to maintain those over a long timeframe."
Here's where things get real. In a few hours, Microsoft and Meta drop their numbers.
Microsoft is expected to report adjusted EPS of $4.24 on revenue of $87.62 billion — roughly 15% year-over-year growth. Azure is still growing at ~40%, and Copilot adoption remains strong. But Microsoft is the second-weakest Mag 7 stock this year, down 21%. The market wants to hear not just that AI products are selling, but that they're selling profitably. At a P/E of 19x — a significant discount to its 10-year average of 27x — the stock looks cheap. But as Brad Warden at Nomura put it: "They look cheap right now, but when you look forward at potential disruption, they are guilty until proven innocent."
Meta is expected to post EPS of $7.19 on $60.22 billion in revenue. Advertising remains the engine, and AI-driven targeting keeps improving. But the real story is capex: Meta raised its 2026 forecast to $125–$145 billion earlier this year, and the company is pouring resources into its in-house "Iris" AI chip (entering production in September) and a new cloud business called Meta Compute. That's a lot of bets on a lot of tables. At 14x forward earnings versus a 10-year average of 20x, Meta is arguably the most undervalued of the bunch — if you believe the spending pays off.
Here's the uncomfortable truth that Tim Cook understood while Satya Nadella, Mark Zuckerberg, and Andy Jassy were signing purchase orders: you don't actually have to build the AI infrastructure to sell the AI.
Apple's strategy is to let Google, OpenAI, and others spend hundreds of billions building models and data centers, then negotiate distribution deals that put those models on 2 billion active devices. It's the same playbook Apple used with cellular modems and display panels. Own the customer. Rent the infrastructure.
The market is rewarding this approach with a vengeance. Apple is up 23% in 2026 and is now the single biggest points contributor to the S&P 500's 8.3% rise. Meanwhile, Microsoft — the supposed AI leader — is down 21%.
But the narrative has a ceiling. Apple avoided the capex trap, yes. But it also has no custom large language model of its own, no cloud AI business to speak of, and its AI features are powered by competitors. If those competitors start charging more for API access — or if regulators force platform-level changes — Apple's "cheap seat at the AI table" becomes a liability. The question isn't whether Apple's strategy looks smart today (it does), but whether it still looks smart in 2028.
There's a quote buried in the Fortune piece that I can't stop turning over. Willy Lee, a principal at Neostellar Capital, said: "We're in a period where people are inclined to sell off on capex, and Microsoft and Meta and Amazon are all holding hands with Alphabet and jumping in to spend."
Holding hands and jumping in to spend.
That's not an investment strategy. That's peer pressure at a trillion-dollar scale.
The original AI capex thesis went like this: first movers would capture insurmountable advantages. Build the biggest models on the biggest clusters, and everyone else would be forced to rent from you. The spending was a moat.
What the market is now asking — loudly and with real money — is whether the moat is actually a money pit. If Google's cloud revenue can grow 82% and the stock still drops 7%, what kind of growth justifies $205 billion in annual spending? Fifty percent? A hundred? What if the answer is "nothing" — because the real winner of the AI boom isn't the company that spends the most, but the one that spends the smartest?
Tonight, Microsoft and Meta get their turn at the podium. Tomorrow, Apple and Amazon. By Friday, we'll know whether the AI capex reckoning was a one-day panic or the beginning of something much bigger.
If I had to bet? The spending isn't stopping. The arms race logic is too embedded, and no CEO wants to be the one who blinked. But the terms of engagement have shifted. The market has stopped grading on effort and started grading on results.
For three years, Big Tech said: "Trust us, the returns are coming."
Today, the market says: "Show us."
Fortune — "Big Tech earnings slam into a market in revolt over AI spending" by Jeran Wittenstein, Ryan Vlastelica (Bloomberg). https://fortune.com/2026/07/26/big-tech-earnings-meta-microsoft-apple-amazon-market-revolt-ai-spending/
Bloomberg — "Magnificent Seven Lose $797 Billion as AI Skeptics Dump Tech Stocks." https://www.bloomberg.com/news/articles/2026-07-23/magnificent-7-loses-767-billion-as-ai-skeptics-dump-tech-stocks
TipRanks — "MSFT, AMZN, META, AAPL Earnings Preview – Here's What to Expect" by Sheryl Sheth. https://www.tipranks.com/news/msft-amzn-meta-aapl-will-report-results-this-week-heres-what-to-expect
Tom's Hardware — "China begins mass production of homegrown immersion chipmaking machines in major breakthrough." https://www.tomshardware.com/tech-industry/semiconductors/china-begins-mass-production-of-domestic-immersion-duv-lithography-machines
Benzinga — "Big Tech Earnings Week: Can AI Spending Finally Deliver Returns?" https://www.benzinga.com/Opinion/26/07/60738619/apple-microsoft-meta-amazon-earnings-why-the-next-48-hours-could-redefine-the-ai-trade
The Motley Fool — "Apple Is Barely Spending on AI While Its Rivals Pour In $700 Billion. Is That Discipline or Denial?" https://www.fool.com/investing/2026/07/27/apple-is-barely-spending-on-ai-while-its-rivals-po/
Implicator — "Magnificent Seven Loses $797 Billion in AI Capex Selloff." https://www.implicator.ai/magnificent-seven-loses-797-billion-after-alphabet-and-tesla-raise-ai-spending/
ValueAdd VC — "Big Tech's $725B AI Spending Tracker (2026)." https://valueaddvc.com/ai-spending
All claims verified against Gold-tier (official filings, analyst estimates via Bloomberg consensus) and Silver-tier (Fortune, Tom's Hardware, TipRanks, Benzinga, Bloomberg) sources. Each source URL was scraped or search-verified. The earnings projections for MSFT/META/AAPL/AMZN are analyst consensus estimates — actual results will be released after market close July 29 (MSFT, META) and July 30 (AAPL, AMZN). Last verified: July 29, 2026.
