Lagarde: AI Stock Valuations Still Excessive, Market Correction Inevitable
nashnova research
ECB President Lagarde warned AI stock valuations remain "very aggressive" despite a recent pullback, with further correction inevitable — while Blackstone's Jon Gray countered the same day that bubble fears are "overstated."
What exactly did Lagarde say?
At the NYT DealBook Summit, Lagarde said AI stocks have already seen a "significant price correction" — but current valuations remain "very aggressive."
Her core test: whether prices can hold depends on companies actually delivering the earnings the market expects, and what those future earnings are worth in today's money.
This means → The ECB stopped short of calling it a "bubble," but made clear this valuation regime is unlikely to last — the wording hedged, the direction did not.
What risks did the ECB report flag?
The ECB's Financial Stability Review, released the same day, identified AI asset valuations, rising corporate debt financing, and geopolitical tensions as simultaneous pressures on financial stability.
The report singled out AI infrastructure's heavy reliance on debt financing: data centers, chips, and cloud infrastructure keep expanding, and tech companies keep borrowing more to fund it.
In plain terms = Much of the AI buildout is being financed with borrowed money. If expectations for AI profitability shift — or earnings fall short of what valuations imply — the price correction could spread from equities into the bond market.
The report also flagged that falling AI model costs could reshape competitive dynamics. This reflects a risk beyond "prices are too high" — the incumbents' market positions themselves may not be secure.
Why is Blackstone pushing back?
Blackstone President Jon Gray said at the TomorrowX conference that bubble fears around AI are "overstated" and that AI infrastructure supply still lags far behind demand.
Gray acknowledged that large-scale technological shifts carry elevated risk, but argued "sentiment has swung too far negative."
This means → Blackstone is not denying risk exists. It is arguing the market has already over-priced the bad news — the real disagreement is not about whether risk exists, but whether demand fundamentals are strong enough to support current spending.
How much has Blackstone itself bet?
Blackstone currently manages roughly $1.4 trillion in assets. AI-related holdings account for about 7% of that.
The firm has assembled a data-center portfolio worth approximately $200 billion. Gray predicted a global consolidation wave, with a handful of players with scale and capital ultimately winning out.
In plain terms = When Gray says fears are "overstated," that is not a neutral observation — Blackstone is one of the largest buyers of AI infrastructure on the planet. His call and his portfolio point in the same direction.
What is the core variable that settles this debate?
Lagarde is watching the gap between valuations and actual earnings delivery — if profits cannot catch up to what stock prices imply, correction is a matter of timing.
Gray is watching demand fundamentals — as long as real-world demand for AI compute keeps growing, the current infrastructure buildout is not a bubble.
This means → The two are not debating whether AI has value. They are debating whether today's prices have already borrowed too much from the future — and the answer depends on whether tech earnings over the next few quarters can deliver real money, not just projections.
AI valuations: correction ahead, or panic overdone?
BULL
Demand outstrips supply
Gray argues AI infrastructure supply still lags demand — the buildout has fundamental backing.
Consolidation favors scale
Global markets will consolidate; players with scale and capital win.
Sentiment already too negative
Gray judges bubble fears "overstated" — the market has over-priced the downside.
BEAR
Valuations still aggressive
Lagarde explicitly called current valuations "very aggressive" — unlikely to persist.
Debt contagion risk
AI expansion runs on borrowed money; earnings misses could trigger a chain reaction into bonds.
Competitive map may be redrawn
Falling AI model costs could reshape the industry — incumbents' positions are not guaranteed.
In plain terms = neither side denies AI's long-term value. The dispute is whether today's prices have already borrowed too much from the future — only actual earnings data can settle it.
市场有风险,内容仅供研究参考,不构成投资建议。
