Apple passed Nvidia on July 17 to reclaim the top spot in global market cap, up roughly 22% in 2026, while Alphabet fell 7% and Meta is down nearly 10% this year after both companies shocked markets with capex guidance that stretches the limits of credibility.
Here's the number that makes the whole thing stark: Apple spent $12.7 billion on capital expenditures in fiscal 2025. Amazon, Alphabet, Meta, and Microsoft, taken together, burned through $416 billion in the same period. And yet on July 17, Apple's market cap climbed past Nvidia's to reclaim the title of the world's most valuable public company, a position it hadn't held since Nvidia seized it in June 2025.
That's not a coincidence. It's a verdict.
The week that delivered Apple its crown was also the week that punished the spenders most visibly. Alphabet reported Q2 2026 results on July 22 that beat estimates on nearly every operational line: total revenue hit $119.8 billion, up 24% year over year, and Google Cloud accelerated to $24.77 billion, up 82%. None of it mattered. What the market saw was $44.9 billion in quarterly capex, free cash flow going negative at -$5.9 billion, and full-year 2026 capex guidance raised to as high as $205 billion. The stock dropped 7%. Meta, which raised its own 2026 spending forecast to between $125 billion and $145 billion after Q1, is down nearly 10% for the year. Investors are watching the cash burn and asking a question the companies haven't answered cleanly: when does this end?
Apple's position here wasn't built by accident. While the rest of the Magnificent Seven treated AI infrastructure as a land grab requiring maximum speed and maximum dollars, Apple quietly leaned on outside compute partners rather than building its own data center empire. Finance chief Kevan Parekh explained on Apple's Q4 2025 earnings call that the company buys computing capacity externally rather than racing to own it. R&D spending has climbed, crossing 10% of revenue for the first time in at least 30 years as of the March quarter, according to CNBC, but capex itself stayed tight. The $4.3 billion Apple spent on capex in the first six months of fiscal 2026 is a rounding error against what Alphabet spent in a single quarter.
The result is free cash flow of $129 billion on a trailing basis. Baird analyst William Power, who raised his price target to $330 on an Outperform rating ahead of Apple's July 30 earnings, called the company a "port in the storm" for software and megacap tech investors spooked by capex escalation elsewhere. That phrase captures exactly what the market is pricing in: Apple as the one large-cap technology company where you don't have to worry about the next spending announcement vaporizing your returns.
Apple's stock has gained nearly 23% in 2026. Nvidia, which held the world's most valuable company title for over a year, is up just 7% over the same stretch. The rotation reflects something real about where investor anxiety is concentrated right now.
Was massive capex ever the right bet? #
That's the question the Apple story is forcing into the open. The prevailing assumption through 2024 and into 2025 was that AI infrastructure spending was table stakes for any technology company that wanted to compete. You built the data centers, trained the frontier models, owned the GPUs, or you fell behind. Alphabet, Meta, Amazon, and Microsoft all internalized that logic and ran with it.
Apple didn't. It partnered with OpenAI for Apple Intelligence features on device, kept its own AI work focused on on-device inference rather than frontier model training, and tied whatever AI gains it could claim directly to its services business and its device upgrade cycle. No AGI moonshot. No hyperscale buildout. Just a bet that its installed base of roughly 2.2 billion active devices was a distribution advantage no data center spend could replicate quickly.
That bet is winning right now, at least on a market cap basis. Whether it holds depends on things that haven't resolved yet. Apple Intelligence hasn't delivered the kind of breakthrough feature that forces a hardware upgrade cycle, and the company's Q3 2025 earnings arriving July 30 will test whether services revenue is actually accelerating in the way the stock's valuation implies. There are also real questions about whether Apple's reliance on partners like OpenAI is durable, or whether it creates dependency risk that will look uncomfortable later.
Frankly, the risk cuts both ways. The companies burning through capex at $205 billion a year are making a bet that the infrastructure they're building will generate returns that justify it. If AI monetization accelerates sharply and the compute they own turns into a structural advantage, today's free cash flow skepticism will look naive. But if the returns stay lumpy and the spending keeps rising, the market's current preference for Apple's model won't look like a temporary rotation. It'll look like a correct read on which companies understood what they were actually building and which ones were just spending to keep up.
For now, $4.9 trillion says the frugal bet is winning. Also read: Nvidia bets $1.5 billion on Amkor to break its chip packaging bottleneck • TSMC beat every earnings record and Wall Street sold the stock anyway • Waymo tells Uber it's going solo in Austin and Atlanta when their contract expires in 2028