Alphabet's Ruth Porat Says 2027 AI Spending Will Rise 'Significantly,' Rejects Jefferies Warning of Capital Destruction

Since Jefferies' Chris Wood published his Friday, Oct. 9 note arguing that cheaper Chinese open-source models will drive massive capital destruction in U.S. AI, Alphabet has answered on the record. On Monday, Oct. 12, Alphabet President and Chief Investment Officer Ruth Porat told the Fortune Most Powerful Women Summit in Santa Barbara that capital spending will "increase significantly" in 2027.

That is on top of a big number already. Alphabet guided to $195 billion to $205 billion of capital expenditures this year, more than double the $91.4 billion it spent in 2025.

The spending

Analyst estimates compiled by Bloomberg put combined 2026 capex at Amazon, Alphabet, Microsoft and Meta above $720 billion, most of it on AI infrastructure.

Porat was asked whether the industry has peaked. She said the buildout is still early relative to the economic value Alphabet expects AI to produce. "We do see this as an opportunity that extends for quite some time," she said.

Alphabet raised its 2026 capex outlook to as much as $205 billion in July. Its shares fell more than 6% the next day.

The strain on the balance sheet

Alphabet spent $44.9 billion on capex in the second quarter alone. Free cash flow was negative $5.9 billion, the first negative quarter since the company's 2004 IPO.

Long-term debt was about $98 billion at the end of June, more than four times the year before.

Alphabet is not the only borrower. Outstanding private loans to AI-related companies have climbed past $200 billion by some estimates, with lenders such as Blue Owl, Apollo and KKR active. Much of that money goes to special purpose vehicles and smaller "neocloud" operators, not to Big Tech itself.

What Porat said back

Porat called Wood's thesis "illogical," pointing to Alphabet's own capacity constraints. She said demand from businesses, government agencies and Alphabet's own platforms is outrunning the computing power it can supply. Profitability, she said, depends on delivering value to partners while providing compute efficiently.

The cloud numbers back her up so far. Google Cloud revenue rose 82% to $24.8 billion in the second quarter, and the unit's operating profit more than tripled to $8.8 billion.

Porat also acknowledged a political problem of the company's own making. She cited a Gallup poll finding 71% of Americans do not want a data center in their community. Her answer was investment in nuclear, geothermal, solar and wind power, plus a skilled-trades training alliance with Ford, BlackRock and Carhartt.

What Wood actually argued

Wood's argument is about pricing and competition, not just demand. He wrote that cheaper Chinese open-source models threaten to commoditize large language models and erode the economics behind American AI spending. He also said China "has become a technological peer to the U.S. in AI."

There is usage data pointing in his direction. Moonshot AI, a leading Chinese open-weight developer, told Fortune that 70% of its usage comes from outside China. New financial figures from Anthropic this week suggest its growth may have flattened, according to Fortune.

Porat's capacity-constraint answer speaks to how much compute customers want. Wood's claim is about what they will pay for models. Those are different questions, and Alphabet has not publicly addressed the second.

One independent analysis, from AMO Research, reads Wood's warning as mainly about financing. It argues the risk is hyperscalers paying for data centers with debt instead of cash, and that the trade could unravel if credit markets stop lending. That analysis also notes Wood's note gave no timeline, and that AI capex was still accelerating when he last spoke.

The Fed's view

The central bank is not united on how to read the boom. Fed Chairman Kevin Warsh said at Jackson Hole in August that more than half of this year's capital-expenditure growth could be attributed to AI, and called business investment the "seed corn of future economic growth."

Governor Christopher Waller said at a Sept. 3 Reuters event that AI investment is a "legitimate part of GDP." Governor Lisa Cook, in a more recent speech, gave greater attention to inflationary pressures that could arrive before any productivity gains.

What is still open

Alphabet can carry this better than most. Fortune's own assessment is that the bigger risk for Alphabet may be not spending, while less well-capitalized players are more exposed.

Porat has not put a dollar figure on 2027. The open question is how much of the industry's next round of spending rides on borrowed money, and what happens to the private lenders and neocloud operators if open-weight competition pushes prices down before the capacity is paid for.