The $1 Trillion Receipt

in #articleyesterday

The $1 Trillion Receipt

Friday was a bill coming due. The market had been running a tab for months — on perpetual AI momentum, on rate cuts that were always three meetings away, on a nine-week winning streak that felt less like fundamentals and more like inertia — and on June 6th, somebody finally asked for the check.

The Philadelphia Semiconductor Index suffered its largest single-day collapse since March 2020. Not a bad week. One day. The S&P 500 dropped 2.64% to 7,383, snapping a nine-week Friday-to-Friday winning streak — the longest since late 2023. The Nasdaq fell 4.18% to 25,709, its worst session since April 2025. Over $1 trillion in semiconductor market cap vaporized in the time it takes to hold a conference call.

The proximate cause was Broadcom. The distal cause was something far less comfortable to name.


Broadcom (AVGO) reported a genuinely strong quarter on Wednesday. Revenue of $22.19 billion, up 48% year over year. AI semiconductor revenue of $10.8 billion — 143% growth, and nearly half the company's total take. Earnings per share of $2.44 against an estimate of $2.39. By any sensible reading, a solid result. The stock was down roughly 15% by Thursday's open.

The mechanism was familiar and increasingly tedious: Hock Tan guided Q3 AI chip sales to $16 billion. Analysts had penciled in $17.2 billion. The company reiterated — rather than raised — its full-year AI semiconductor forecast of "in excess of $100 billion." It also quietly walked back its earlier ambition to supply integrated AI systems to hyperscalers, pivoting to "chips only." CEO Tan noted that anchor client Google may diversify its custom silicon supply chain going forward.

None of this is catastrophic. All of it is meaningful.

What the market was actually doing — underneath the algorithmic panic selling in Micron (MU, -13%), AMD (-11%), Intel (-11%), and Marvell (MRVL, -16%) — was revising an assumption it had held too long and too loosely: that AI capex would accelerate forever, that the hyperscalers would spend without ceiling, and that any company sitting in the critical path of that infrastructure build was exempt from ordinary financial physics. A guidance miss of $1.2 billion on a quarterly number became the occasion to ask whether the entire thesis was priced correctly. The answer, apparently, was no.

Contagion crossed the Pacific by morning. South Korea's KOSPI fell 5.54%. Samsung Electronics dropped 6.4%. SK Hynix — the memory supplier that rode the HBM wave to dizzying heights — shed nearly 10% in a session. Dutch lithography supplier ASML slipped 3.8%. The geography of the damage was a map of the AI supply chain itself.


Then the jobs report landed and made everything worse.

The U.S. economy added 172,000 nonfarm payrolls in May — more than double the consensus estimate of around 82,000. The unemployment rate held at 4.3% for a third straight month. Prior months were revised upward by a combined 93,000. This was the third straight labor print that destroyed the prevailing narrative: that hiring was cooling, that the Fed would eventually blink, that patience had a payout.

It doesn't. Not anymore.

Within minutes of the BLS release, CME FedWatch flipped. Odds of at least one Federal Reserve rate hike in 2026 crossed above 57% — higher than the combined probability of a hold or a cut. Bond traders, per Bloomberg, fully priced in a hike by year-end. The ten-year Treasury yield ripped higher. Bitcoin fell more than 5% to below $60,000. Gold dropped 3.5%. Everything with a duration premium and a risk premium sold simultaneously, which is the market's least subtle way of saying it's scared.

Kevin Warsh chairs the Fed for the first time on June 17th. Whatever you think of his appointment — and the political circumstances that produced it — his first meeting now arrives with the bond market pricing a hike, an AI sector in freefall, a jobs market that refuses to cooperate with the soft-landing narrative, and energy inflation that the ceasefire optimism of April already proved fragile. It is, to put it politely, a consequential debut.


The deeper issue is structural, and it's been visible for months to anyone willing to squint past the Nasdaq records.

Eight of eleven S&P 500 sectors finished May in negative territory. The technology sector accounted for essentially all of the index's 5.26% monthly gain. Everything else — energy, utilities, staples, financials, industrials — was being quietly liquidated to fund the chip trade. The Russell 3000 Growth index outperformed Value by over four percentage points. Long semiconductors polled as the most crowded trade on Wall Street at 73% of survey respondents. That's not a consensus. That's a fire door with one exit.

When the most crowded trade in the market meets a guidance miss and a jobs report that puts hikes back on the table, the exit gets very narrow very fast. Friday was the test of the architecture. The architecture failed it.

What's interesting now isn't whether AI spending is real — it is — or whether Broadcom's business is impaired — it isn't, revenue is still compounding at nearly 50% year over year. What's interesting is whether the market can hold together a bull thesis built on perpetual expansion while the Fed is being backed into a corner by data that says the economy is too hot, not too cold. Rate cuts were the lubricant for the whole AI supercycle narrative. Remove them — or replace them with the prospect of hikes — and valuations that made sense at 4% ten-year yields look different at 4.6% and climbing.

The Dow closed Thursday at a fresh record. UnitedHealth, JPMorgan, Walmart led the charge — value, not growth, not chips, not AI. The rotation was visible in real time: money leaving Nvidia (NVDA, -5.93% on Friday) and arriving at companies with predictable cash flows and manageable duration risk. It may be premature to call it a regime change. But when the Dow sets records while the Nasdaq loses four percent in a single session, something structural is being communicated.

The market ran a nine-week tab. Friday was the receipt.

The question now is whether the kitchen is still open.

Sort:  

Upvoted! Thank you for supporting witness @jswit.

Your in-depth analysis of the market's sudden shift is really insightful, especially with the specific details on Broadcom's report - that 48% revenue growth is quite impressive. 💸📊