SPECIAL REPORTBEARISH

FT Survey Finds Cracks Forming in $10 Trillion AI Buildout

Financial Times reporting points to physical-world constraints — power, land, hardware supply — colliding with the scale of planned AI data center spending.

Executive takeaway

The Financial Times reports that the roughly $10 trillion bet on AI infrastructure is running into the practical limits of electricity supply, land and physical build-out speed.

Newsroom graphic
CoreWeave diverges from broader AI trade. Plotting CoreWeave against a semiconductor/AI benchmark shows whether the financing stress flagged in its downgrade is already visible in its stock price relative to peers riding the same buildout.

CoreWeave diverges from broader AI trade

Plotting CoreWeave against a semiconductor/AI benchmark shows whether the financing stress flagged in its downgrade is already visible in its stock price relative to peers riding the same buildout.

Live price data for CRWV and SMH, referenced alongside the Seeking Alpha CoreWeave rating downgrade.

Use this chart

Free to embed with attribution:

<figure><a href="https://www.indy.finance/news/ft-survey-finds-cracks-forming-in-10-trillion-ai-buildout"><img src="https://www.indy.finance/news/ft-survey-finds-cracks-forming-in-10-trillion-ai-buildout/graphic.svg" alt="CoreWeave diverges from broader AI trade" width="1200" height="675"></a><figcaption>CoreWeave diverges from broader AI trade — <a href="https://www.indy.finance/news/ft-survey-finds-cracks-forming-in-10-trillion-ai-buildout">Indy Finance</a></figcaption></figure>
A Financial Times report describes a widening gap between the pace of AI infrastructure announcements and the physical capacity needed to build them. The piece frames the problem as one of unforgiving physical-world economics: power generation, grid connections and construction timelines do not scale as fast as capital commitments. The report lands alongside a separate Seeking Alpha downgrade of CoreWeave, a cloud computing firm built around AI workloads, which argues the company can survive but that shareholders could still lose money — evidence that financing stress is starting to show up in individual names tied to the AI buildout, not just in macro commentary. What remains unresolved is how much of the roughly $10 trillion figure is committed capital versus announced intent, and how quickly power and permitting bottlenecks could ease or worsen over the next several quarters.
What would change this view

If major hyperscalers report in coming quarterly earnings that data center capex guidance is being raised rather than trimmed, it would undercut the fraying-boom framing.

Wire sources cited

Produced automatically by the INDY NEWS Desk from the public wire sources cited above, and checked against them before publication. Market commentary, not investment advice.