Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/THE AI CORNER/The AI trade just lost $1 trilli…
NEWS
// NEWSLETTER ISSUE
THE AI CORNER

The AI trade just lost $1 trillion. Here is where the value went

DATE July 17, 2026SOURCE THE AI CORNERPARTICIPANTS THE AI CORNER
In this episode
// SUMMARY

1. Key Themes


Theme 1: Sentiment-Driven Selloff Decoupled from Accelerating Fundamentals

The semiconductor sector lost over $1 trillion in market cap while underlying demand data was simultaneously hitting record highs — a rare and meaningful dislocation.

"Here is what makes this moment interesting: the demand data kept getting stronger the whole way down."

Supporting data points cited in the article:

  • Hyperscaler capex tracking ~$725B for 2026, up 77% from $410B
  • Goldman projecting big-tech capex surpassing $1 trillion in 2027
  • TSMC raised full-year guidance to 40%+ growth and lifted capex to $60–64B
  • Micron printed revenue up 346% and a record 84.6% gross margin before its stock bottomed

Theme 2: Dislocations Create Both Opportunities and Value Traps

The author's central thesis is that not all drawdowns are equal — some are genuine mispricings, others are cheap-looking stocks with structurally broken stories.

"When prices fall 30-48% while fundamentals accelerate, you get dislocations. Some are gifts. Some are traps wearing a low multiple. Telling them apart is the entire game."


Theme 3: AI Infrastructure Capex Cycle Is Still Early-Innings

Analyst commentary frames this pullback not as a cycle peak but as a mid-game pause in a long structural buildout.

"Wedbush's Dan Ives called the cycle '3rd inning, 1 out in a 9-inning game.' Jefferies' Brent Thill called the bear thesis 'garbage.'"


Theme 4: Macro and Sentiment Catalysts Overwhelmed Fundamentals Temporarily

The selloff was driven by a confluence of non-fundamental factors, not deteriorating business conditions.

"The sell-off ran on sentiment: Meta's plan to sell surplus compute, a hawkish new Fed chair, 4.48% yields, memory 'peak' warnings, and Michael Burry's disclosed shorts."


2. Contrarian Perspectives


Perspective 1: Memory "Peak" Warnings Are Premature

The bear case on memory — particularly Micron — centered on peak cycle fears. But Micron actually guided higher even as its stock was collapsing 30%.

"Micron printed revenue up 346%, a record 84.6% gross margin, and guided higher, three weeks before the stock bottomed."

This directly contradicts the peak-memory narrative that drove the selloff. The evidence suggests the warnings were sentiment-driven, not data-driven.


Perspective 2: The AI Trade Is Being Mis-Priced as a Bubble When Capex Data Suggests Structural Demand

The consensus fear is that AI infrastructure spending is speculative and unsustainable. However, committed hyperscaler capex of $725B in 2026 (up 77% YoY) and a Goldman forecast of $1 trillion by 2027 suggest the opposite — this is contracted, scaling demand.

"The four largest hyperscalers are tracking ~$725B of 2026 capex, up 77% from $410B last year. Goldman now expects big-tech capex to pass $1 trillion in 2027."


Perspective 3: Meta Selling Surplus Compute Is a Bearish Signal — But May Be Misleading

The market treated Meta's plan to sell surplus compute as a demand destruction signal. However, surplus compute at one hyperscaler does not indicate broad demand deterioration when aggregate capex is at record highs — it may simply reflect reallocation or capacity management.

"The sell-off ran on sentiment: Meta's plan to sell surplus compute... [was one of the drivers]."

The article implicitly frames this as a sentiment overreaction rather than a structural shift, given the simultaneous acceleration in aggregate capex data.


3. Companies Identified

CompanyDescriptionWhy MentionedQuote
NvidiaLeading AI chip designerLost ~$1 trillion in market cap during the selloff; flagged as a key dislocated name"Nvidia shed a trillion on its own."
MicronMemory semiconductor manufacturerDropped 30% in three weeks despite record fundamentals; highlighted as a potential deep-value setup"Micron printed revenue up 346%, a record 84.6% gross margin, and guided higher, three weeks before the stock bottomed."
CoreWeaveAI cloud infrastructure / neocloudNearly halved in price; mentioned as part of the broader selloff"CoreWeave nearly halved."
TSMCWorld's leading semiconductor foundryBeat earnings, raised full-year guidance to 40%+ growth, lifted capex to $60–64B"TSMC just beat, raised full-year guidance to 40%+ growth, and lifted capex to $60-64B, this week."
MetaSocial media / AI hyperscalerIts plan to sell surplus compute was cited as a sentiment trigger for the selloff"The sell-off ran on sentiment: Meta's plan to sell surplus compute..."

4. People Identified

PersonDescriptionWhy MentionedQuote
Dan IvesSenior analyst, Wedbush SecuritiesProvided bullish framing on the AI cycle's duration"Wedbush's Dan Ives called the cycle '3rd inning, 1 out in a 9-inning game.'"
Brent ThillSenior analyst, JefferiesDismissed the bear thesis on AI semiconductors outright"Jefferies' Brent Thill called the bear thesis 'garbage.'"
Michael BurryInvestor, Scion Asset Management; known for The Big ShortHis disclosed short positions contributed to negative sentiment in the sector"...and Michael Burry's disclosed shorts."
Ruben DominguezAuthor, The AI Corner newsletterBuilt the analytical framework (Pullback Playbook) for navigating the dislocationByline author of the article

5. Operating Insights


Insight 1: Build a Threshold System, Not a Static View

The article advocates for reacting to evidence rather than headlines by defining in advance the specific data points that would change a thesis — converting a stock from "value" to "trap" (or vice versa).

"The threshold system — the exact data points that flip each thesis from value to trap, so you react to evidence instead of headlines."

For operators and investors, this is a discipline of pre-committing to decision criteria so that price moves don't trigger emotional reactions.


Insight 2: Structure Positions Across Core, Satellite, and Trading Sleeves

Rather than treating the pullback as binary (buy or avoid), the article frames position-sizing as a tiered exercise — separating high-conviction core holdings from more speculative satellite and trading positions.

"The sizing frame — how to structure core, satellite, and trading sleeves against the one number that keeps bulls honest."


6. Overlooked Insights


Insight 1: The VanEck Semiconductor ETF as a Macro Sentiment Gauge

The article cites the ETF dropping 5% in a single session as a marker for the sector-wide repricing. For investors, this suggests using the ETF's single-day moves as a real-time sentiment signal — a 5%+ intraday drop may indicate sentiment-driven (rather than fundamental-driven) selling that warrants closer scrutiny of individual name fundamentals.

"The VanEck Semiconductor ETF dropped 5% in a single session, and the sector lost over $1 trillion in market value in under two months."


Insight 2: Calendar-Driven Thesis Resolution

The article alludes to three specific upcoming dates within six weeks that the author believes will "settle the argument" on the bull vs. bear debate. This implies that for active investors, the near-term data calendar (likely earnings or capex guidance events) should be a primary input into position timing — not just valuation.

"The calendar — the three dates in the next six weeks that settle the argument."

These dates are paywalled, but the framework itself — identifying binary event dates and sizing positions accordingly — is a transferable tactical discipline.