Dan Gallagher
The AI spending race among the world's largest companies is now in its third year -- with no end in sight.
When Microsoft, Amazon, Meta Platforms and Google-parent Alphabet report their quarterly results on the same afternoon next week, investors will scrutinize the capital expenditures line item. And with good reason. Those four companies spent a combined $410 billion on capex last year, which is nearly triple what they spent in 2022.
The launch of ChatGPT late that year sparked a technological arms race. Four of the world's largest companies by revenue and market capitalization suddenly feared that falling behind in AI meant early obsolescence. That fear is still a driving force; Amazon, Meta and Google's parent all used their quarterly reports three months ago to project big jumps in capital spending this year.
Microsoft remains the wild card. Because the software giant has a fiscal year that ends in June, it typically doesn't forecast for calendar-year periods. But Wall Street analysts still expect a 31% jump in Microsoft's total capital spending for 2026, according to estimates from Visible Alpha. That would make for as much as $674 billion in total capital spending for the four companies this year. And it would mark the third year in a row that total combined spending growth has exceeded 60%.
Even the deepest pockets in the world can't sustain that level of spending for long. Unless, that is, AI starts to drive significant revenue and earnings growth.
The rub is that pulling back now could also look like an admission of weakness -- especially with AI model makers OpenAI and Anthropic sharply growing their own revenue ahead of expected public offerings later this year. That gives the big tech companies little incentive to tap the spending brakes.
If their AI businesses do start showing noticeable growth, the big tech companies will still face major costs down the road. Large capital expenditures mean large depreciation expenses in the future. Big-ticket items like Nvidia's chips are fully written down over roughly five-year periods.
Depreciation charges for Google -- considered the AI front-runner of the group -- are expected to total 35% of net income in 2028 compared with an average of 17% over the past five years. The bills for big tech's spending spree will last a long while.
This is an edition of the WSJ AI & Business newsletter, a weekly digest to help you make sense of AI's impact on business with news, insights and data from our global team of technology journalists. If you're not subscribed, sign up here.
Amazon Ups its Bet on Anthropic
Amazon is investing another $5 billion in Anthropic, which could scale up to $25 billion if certain commercial milestones are met. In return, Anthropic has agreed to purchase more than $100 billion worth of Amazon's cloud services. The deal deepens the relationship between the AI startup that is challenging OpenAI for supremacy in the business of building frontier models, and world's largest provider of cloud computing services.
The Number
OpenAI chief Sam Altman's salary in 2024, the most recent year for which data is available. Altman's more lucrative ties to companies OpenAI has done business with are raising questions ahead of the AI darling's IPO debut likely this year or next.
What the Humans Are Saying
AI in Charts
Intel has gone from stock-market laggard to industry outperformer over the past few months, due largely to a recent chip-development deal with AI titan Nvidia. A shift in AI computing toward deploying AI agents -- an area where Intel's central processing chips excel -- has also played into Intel's hands.
While that enthusiasm has pushed Intel's valuation to levels not seen in decades, the storied American chip maker still has a long way to go before it recovers its mojo. Once the undisputed leader in chip design and manufacturing, Intel remains behind competitors in both areas, and prospects for a resurgence haven't meaningfully improved lately. Intel also has yet to build a large and successful contract chip-making business that would likely be necessary if its financial performance is to match investors' increasingly lofty expectations.
AI in the Wild
Software giant Adobe is proffering an AI agent for corporate customers that it's hoping can help revive a business that's been on the ropes lately. Shares of Adobe, which makes software widely used in video and image editing, have fallen more than 25% this year amid growing doubts about the software-as-a-service business model in general and Adobe in particular.
Other Highlights From the Week in AI
-- Anthropic CEO Dario Amodei met with President Trump on Friday to try to
thaw tensions centering around Anthropic's refusal to give the
administration free rein to use its software in weaponry and spying.
-- Those challenges aside, Anthropic's business is growing at a healthy
pace: The company is expanding operations in the U.K., leasing space for
up to 800 employees in London, quadruple its current footprint there.
-- Cerebras Systems, a startup that makes gigantic chips for AI calculations,
is planning to go public after reaching big financial deals with OpenAI
and Amazon's cloud-computing operation.
-- Corporate leaders often demur on the job-eliminating potential of AI. Not
Verizon's Dan Schulman. The carrier's chief executive says AI could cause
the unemployment rate to rise to 20% to 30% in the next several years.
About Us
WSJ AI & Business is a weekly look at AI's transformation of the business world. This newsletter was curated and edited by Dan Gallagher and Asa Fitch. Reach them at dan.gallagher@wsj.com and asa.fitch@wsj.com (if you're reading this in your inbox, you can just hit reply). Got a tip for us? Here's how to submit.
(END) Dow Jones Newswires
April 21, 2026 12:10 ET (16:10 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments