BlackRock Inc.
"We're still in massive build-out phase," she said.
Big Tech Still Can't Build Fast Enough
The clearest evidence for Li's "massive build-out" phase is that the biggest spenders still say capacity is falling short of demand.
Amazon, Alphabet and Microsoft are on track to spend roughly $590 billion to $600 billion on capital expenditures this year, much of it tied to data centers and AI infrastructure.
Amazon raised its 2026 capex forecast to about $220 billion, yet CEO Andy Jassy has said even that may not provide enough capacity to meet demand. Alphabet lifted its forecast to $195 billion toi $205 billion while saying demand continues to exceed available computing capacity.
Microsoft expects about $175 billion of calendar-year 2026 capex, with CFO Amy Hood similarly saying customer demand "continues to exceed available capacity."
Nvidia
Asked whether AI is a productivity revolution or a bubble, Li said: "It can be both."
For now, however, she doesn't think it's a bubble. Li said tech valuations aren't particularly stretched despite companies increasing capex and "doubling down" on AI because they are still doing the "heavy lifting in earnings delivery."
Traders Put Low Odds on an AI Bust
Prediction market traders aren't heavily betting on an imminent collapse either.
Polymarket gives an AI industry downturn by Dec. 31 roughly a 9% chance, with about $2.35 million in volume on that contract.
The bar is high. At least three severe events must occur within 90 days, including possibilities such as Nvidia falling 50% from its all-time high, the SOXX semiconductor ETF falling 40%, or OpenAI or Anthropic declaring bankruptcy.
Li's concern is less dramatic. She warned that "narratives can get ahead of actual revenue creation and adoption."
BlackRock Is Already Using AI
Li said the firm uses outside LLMs and its own tools, including Asimov, an internal investment platform, and is looking to integrate AI into "every step of our day-to-day."
But while BlackRock sees productivity gains in some companies and sectors, Li said the broader economy has "yet to see meaningful progress."
Despite the hundreds of billions already being spent, Li still sees AI in the build-out phase, with broader adoption and productivity gains yet to follow.
