Analysis
Four days after Jefferies' Chris Wood told Bloomberg that the most likely long-term outcome of the AI boom is large-scale U.S. capital destruction, Alphabet president and chief investment officer Ruth Porat pushed back directly, telling Fortune it's 'illogical to say we're in a capacity-constrained position' and that the AI buildout's economic value is still underpriced.
What's changed since Pulse covered Porat defending Alphabet's AI spending earlier this month:
- New capex guidance -- $195B-$205B for 2026; Porat now says Alphabet will spend 'significantly more' in 2027.
- New argument from Wood -- cheaper Chinese open-weight models, like Moonshot AI (which he says draws roughly 70% of its usage from outside China), could commoditize large language models faster than closed-model labs such as OpenAI, Anthropic and Google Gemini can defend their pricing.
“Alphabet's 2025 capex was $91.4 billion, so the new guidance is more than double that.”
Alphabet's 2025 capex was $91.4 billion, so the new guidance is more than double that.
The numbers Porat isn't disputing are the ones that make Wood's case for him:
- Free cash flow -- Alphabet posted negative free cash flow in Q2, its first since its 2004 IPO.
- Long-term debt -- about $98B at the end of June, more than four times a year earlier.
- Big Tech capex -- Amazon, Alphabet, Microsoft and Meta are together on pace for more than $720B in 2026 capex, per Bloomberg-compiled estimates.
- Private credit -- Blue Owl, Apollo and KKR have extended over $200B in loans to AI-related borrowers on top of that.
Porat's answer to all of it is essentially a time-horizon argument: that the payoff window is longer than quarterly numbers suggest, and that differentiated AI products will keep paying customers regardless of how good the open-weight alternatives get. That's a bet Alphabet can better absorb the risk of than most of its AI-lab customers can. Wood is now on record saying that bet won't pay off.