Nov 5, 2026 · 43 min
AI sell-off tests the market’s growth story
Closing Bell 11/5/26
The decline is forcing investors to distinguish between a temporary liquidation shock and a deeper reset in AI spending, financing, and valuations.
- 1AI infrastructure spending may peak sooner, with some planned data centers potentially never getting built.
- 2Leverage, margin calls, and financing exposure are spreading the sell-off beyond semiconductors into banks and private credit.
- 3Tom Lee sees policy fears and forced liquidation driving an overreaction, while options levels frame the market’s next test.
Don't miss
Tom Lee argues that policy fears and forced liquidation have produced an overshoot, with AI demand intact despite delayed or relocated construction.
The brief
A broad sell-off hits technology, semiconductors, data-center infrastructure, and NVIDIA, raising fresh doubts about future capacity and the durability of the AI trade.
Dan Ives now expects at least 10% of planned data centers may not be built and sees hyperscaler capital spending peaking in 2027, while remaining bullish on hyperscalers and software.
Anthropic’s planned IPO faces a tougher market: demand remains strong, but pricing could move toward the lower end as its growth story has limited room to improve.
The pullback spreads through banks and private credit as investors reassess data-center loans, infrastructure financing, leverage, and the broader consequences of an AI spending reset.
Tom Lee calls the move a policy and liquidation shock, arguing that AI demand remains intact even as construction shifts and prices overreact to a slowdown rather than a collapse.
The closing panel weighs forced selling, options levels, and circuit-breaker risks while asking whether prices are signaling excessive enthusiasm before earnings estimates catch up.
What was said on this episode
31 statements · 17 positive · 11 negative · 3 neutral
Dan Ives expects 10–20% of planned data centers not to be built
“10% of data centers, and that could be low. It could be 15%, 20% of data centers that ultimately don't happen.”
Listen at 3:04
Big Tech capital expenditure growth will peak in 2027
“our view now is that CapEx for big tech. It peaks out in 2027, next year”
Listen at 3:11
Hyperscalers are best positioned amid constrained AI capacity
“the hyperscalers are in the best position”
Listen at 3:47
AI data-center demand exceeds current supply by 40–50%
“demand still 40 to 50 percent over the supply”
Listen at 4:49
Demand for enterprise AI software projects exceeds available supply fourfold
“demand basically is four to one”
Listen at 6:26
The AI buildout will continue but over a longer timeline
“this AI revolution build-out is not going to happen. It's just going to be more elongated”
Listen at 7:11
NVIDIA earnings estimates are understated by 15–20%
“I think they're underestimated by 15 to 20%”
Listen at 8:08
About 30% of committed AI capacity may be canceled
“30% of that comes off”
Listen at 9:27
Neocloud providers will undergo substantial consolidation
“it's going to be massive consolidation, musical chairs”
Listen at 9:51
Anthropic’s valuation may fall from $1.5 trillion toward $1.2–$1.3 trillion
“$1.5 trillion. Now is it 1.3? Is it 1.2?”
Listen at 10:32
Anthropic’s IPO valuation could decline because compute allocation may fall
“the price, you know, very logically could come down a bit as it goes out”
Listen at 12:39
Thirty-eight jurisdictions voted to restrict data centers
“38 jurisdictions voted to restrict data centers”
Listen at 13:31
AI companies will eventually build the restricted data centers elsewhere
“They will be able to build those data centers eventually”
Listen at 13:42
Data-center location restrictions will slow AI growth and alter competition
“That slows the story. It shakes up the competitive balance.”
Listen at 13:58
Anthropic is unlikely to delay its IPO
“I don't see them actually pushing off.”
Listen at 15:05
Margin leverage reached 125%, versus 75% at recent peaks
“margin was 125 percent. Normally, last couple of peaks, it was 75 percent.”
Listen at 15:49
The AI sell-off is neither a bubble burst nor mania’s end
“This is not the bursting of a bubble. This is not the end of a mania.”
Listen at 16:45
Today’s market decline is an excessive downside move
“today is an overshoot to the downside”
Listen at 17:48
The economy has entered a slowdown within an ongoing cycle
“we've gone into a bit of a slowdown”
Listen at 19:43
Value stocks may outperform growth stocks temporarily
“value may outperform growth for a while”
Listen at 20:26
Robinhood users are buying broad indices and NVIDIA during the sell-off
“they are actually buying like broad-based indices and NVIDIA”
Listen at 20:44
AI demand will not slow because companies continue benefiting
“the demand side for AI isn't going to slow”
Listen at 28:37
The market has overreacted in repricing AI-related stocks
“Does that warrant the massive, massive repricing of these stocks? I think that's overly done.”
Listen at 29:04
The Fed may support markets if AI weakness becomes an employment risk
“the Fed put comes into play if it becomes a employment risk story”
Listen at 30:11
Companies may retrofit vacant industrial and office space into data centers
“you may be retrofitting. empty industrial plants and office space to put in data centers”
Listen at 31:07
The market sell-off is unlikely to end after one day
“I don't think it's a one-day event”
Listen at 35:33
Already permitted data centers will likely proceed without interruption
“Permitted, already permitted data centers are probably, almost surely, going to go on without a hitch.”
Listen at 39:24
Micron could generate $300 billion in free cash over seven to eight quarters
“I still think they're pretty good to do $300 billion in free cash in the next seven, eight quarters.”
Listen at 41:16
Investors should cautiously buy companies already generating earnings
“I'd rather buy the immediate stuff. I would dip my toe there.”
Listen at 41:34
Investors should diversify portfolios across sectors
“you've got to balance it with health care, energy, et cetera. It's always got to be diversified.”
Listen at 41:45
Stock prices often move before underlying fundamentals
“Quite often, prices lead fundamentals.”
Listen at 42:12
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.