This essay seems within the Oct. 17, 2026 version of the Fortune 500 Digest e-newsletter, which rounds up the headlines driving the week’s most vital enterprise information and protection of Fortune 500 corporations. Subscribe to obtain it in your inbox each Saturday morning.
The AI buildout is in full swing, and it’s extremely costly.
Between Amazon, Alphabet, Microsoft, and Meta, over $720 billion of capital expenditures is projected to be spent this yr, most of it on AI infrastructure, in line with analyst estimates compiled by Bloomberg. Alphabet President and Chief Funding Officer Ruth Porat confirmed to me this week on the Fortune Most Highly effective Girls Summit that Alphabet intends to spend “considerably extra” in 2027.
This yr, Alphabet mentioned it expects capital expenditures of $195 billion to $205 billion. That’s greater than double the $91.4 billion it spent in 2025. When requested whether or not spending on the buildout has peaked throughout the business, Porat mentioned it’s nonetheless in its early phases relative to the eventual financial worth Alphabet thinks AI can drive. “We do see this as a possibility that extends for fairly a while,” she mentioned.
More and more, debt is fueling the AI buildout too. Personal credit score funds reminiscent of Blue Owl, Apollo, and KKR are a rising supply: Excellent personal loans to AI-related corporations have climbed to greater than $200 billion by some estimates. Usually, they’re lending to particular objective autos and smaller “neocloud” operators quite than to deep-pocketed Massive Tech gamers. Even Alphabet posted detrimental free money circulation in Q2 for the primary time since its 2004 IPO to assist fund AI. Its long-term debt was about $98 billion on the finish of June, up greater than 4 occasions the yr prior.
Alphabet can afford such pricey long-term AI bets—and arguably the larger threat for it might be not making them. However many much less well-capitalized gamers are susceptible, and up to date shifts could also be making them extra so.
One of many largest: Open-weight fashions, significantly from China, are more and more taking enterprise and utilization from closed fashions like Anthropic, OpenAI, and Google’s Gemini. New monetary figures from Anthropic this week additionally counsel its progress could have flattened.
This development led Jefferies’ international head of fairness technique, Chris Wooden, to declare that the almost definitely long-term end result of the AI growth is huge capital destruction within the U.S., in line with Bloomberg. Wooden argued that cheaper Chinese language open-source fashions threaten to commoditize massive language fashions, eroding the economics behind America’s AI spending, and that China “has develop into a technological peer to the U.S. in AI.”
After I just lately spoke with one in every of China’s main open-weight mannequin builders, Moonshot AI, I realized that 70% of its utilization comes from exterior China.
I requested Porat straight about Wooden’s thesis that a lot of the cash pouring into the U.S. AI race may very well be destroyed.
“It’s illogical to say we’re in a capacity-constrained place as a result of there’s such robust demand for the perception within the capacity to enhance enterprise and authorities, and but you possibly can’t drive a return on it,” Porat replied. She additionally famous that relating to the development towards open-weight fashions versus closed, she feels AI corporations that add differentiated worth will discover clients paying for his or her providers regardless.
Porat has had a front-row seat to earlier bubbles, first as head of Morgan Stanley’s expertise banking workforce in the course of the dot-com period, and later in the course of the 2008 monetary disaster, when she helped advise then U.S. Treasury Secretary Hank Paulson. She went on to develop into Morgan Stanley’s CFO in 2010.
Her takeaway from these intervals is that corporations utilizing AI to construct really differentiated experiences will win.
“One of many key classes from the monetary disaster was that it is best to determine your biggest supply of vulnerability early and fortify towards it, as a result of you possibly can’t handle it within the second. For banks, it was liquidity. With out liquidity, you couldn’t run a buying and selling operation. With out liquidity, you shut down,” Porat mentioned.
“I’ve been requested that query often since: What’s the biggest supply of vulnerability immediately? I believe for all of us throughout business, one of the vital parts is innovation in what you are promoting. If anybody is connecting together with your clients in methods you aren’t, as a result of they’re utilizing AI and you aren’t, you’re at a aggressive drawback.”
Sure, she’s speaking her ebook right here, and the “simply add extra AI” view doesn’t negate the potential systemic financial threat of overbuilding. However it’s additionally true that AI just isn’t going away and the necessity for corporations to adapt is simply going to extend.
At Alphabet, Porat indicated that the corporate will preserve investing closely in AI due to its conviction within the long-term financial upside.
“You will need to make investments for long run progress,” Porat mentioned. “By the point you want it, it’s too late to say, OK, now we’re going to begin.”
For extra, take a look at my full dialog with Google CIO and President Ruth Porat right here.
