Alphabet Will Spend “Significantly More” on AI in 2027 as Jefferies’ Chris Wood Warns of U.S. Capital Destruction
Alphabet plans to spend “significantly more” in 2027 than the $195 billion to $205 billion in capital expenditures it expects this year, President and Chief Investment Officer Ruth Porat told Fortune’s Alyson Shontell at the Fortune Most Powerful Women Summit this week. Capital expenditures, or capex, are the money spent on data centers, chips and other long-lived equipment. This year’s guidance is already more than double the $91.4 billion Alphabet spent in 2025, about 2.2 times at the midpoint.
Porat was answering Jefferies’ global head of equity strategy, Chris Wood, who has said the most likely long-term outcome of the AI boom is massive capital destruction in the U.S., according to Bloomberg as cited by Fortune. Wood argued that cheaper Chinese open-source models threaten to commoditize large language models, the systems behind chatbots, and erode the economics behind America’s AI spending.
“It’s illogical to say we’re in a capacity-constrained position because there’s such strong demand for the insight in the ability to improve business and government, and yet you can’t drive a return on it,” Porat told Fortune.
What Alphabet is carrying
Alphabet posted negative free cash flow in the second quarter, meaning it spent more on operations and new equipment than its business brought in, for the first time since its 2004 IPO, Fortune reports. Its long-term debt was about $98 billion at the end of June, more than four times the year before.
Shontell writes that Alphabet “can afford such costly long-term AI bets,” and that arguably the bigger risk for it would be not making them. Across Amazon, Alphabet, Microsoft and Meta, more than $720 billion in capex is projected this year, most of it for AI infrastructure, according to analyst estimates compiled by Bloomberg, as Fortune reports.
Porat said Alphabet will keep investing heavily because it believes in AI’s long-term economic upside. “You must invest for long term growth,” she said. “By the time you need it, it’s too late to say, OK, now we’re going to start.”
Who borrowed to build
A growing share of the buildout runs on debt. Outstanding private loans to AI-related companies have climbed past $200 billion by some estimates, Fortune reports, from private credit funds such as Blue Owl, Apollo and KKR. Those funds often lend to special purpose vehicles, separate companies set up to hold a single project and its debt, and to smaller “neocloud” operators that rent out AI computing, rather than to Big Tech.
Open-weight models, ones whose files anyone can download and run, are pulling business and usage away from closed models like Google’s Gemini, Fortune reports, particularly models from China. Shontell writes that China’s Moonshot AI told her 70% of its usage comes from outside China.
On that shift, Porat said AI companies that add differentiated value will find customers paying for their services regardless, per Fortune.
Full data centers, cheaper AI
Porat’s answer is that customers want more computing than exists, so the capacity gets used and the spending pays back. Wood’s warning is that a downloadable model doing nearly the same job pulls what anyone can charge for AI toward the cost of running it, and full data centers can still rent out a product that keeps getting cheaper.
The neoclouds those funds often lend to make their money renting out that computing, and the rent is the number Wood’s argument puts at risk.
Sources
Discover more from Bitter Fool
Subscribe to get the latest posts sent to your email.