Inside the Broadcom Rout: When the Market Finally Admits Its Faith Was Misplaced
Friday morning, 9:31 a.m. Eastern. Broadcom opens at its lowest since before earnings were announced. Down 11%. Not down from a previous high—down from the moment it hits your screen. Your portfolio feels it immediately. A cold burn. Not catastrophic, but final. Like you just got told something you suspected but didn't want confirmed.
The stock—AVGO, market cap $750 billion, ninth largest company in the world—had just reported earnings people were supposed to love. Revenue up 13% year-over-year. New customer in Anthropic. The setup was perfect. And it died anyway.
By 10:47 a.m., it's down past 11%. Broadcom has shed more value than most companies are worth. Somewhere in the market, someone's $1.2 million position just became $1.07 million. That's $130,000 gone before lunch. They're still sitting at the coffee maker thinking about whether it's time to update the resume.
The Canary That Stopped Singing
For three days, the AI infrastructure thesis had been sliding. Wednesday you got the Fed rate cut—25 basis points, Powell talking about being "well-positioned to wait," which is central banker speak for "we're done." The market heard it and started asking an uncomfortable question: if we're not getting cheap capital anymore, what are all these data center projects actually for?
Oracle had already telegraphed the problem Thursday. Guided revenue higher. Also guided delays: some data centers that were supposed to open in 2025? Now 2027. Your neural network of confidence in AI capex starts firing warning signals. Something is broken in the chain.
Then Broadcom.
"Backlog and margins commentary was muddled," UBS said Friday morning, which is analyst-speak for "they just admitted gross margins are getting squeezed and nobody knows why they told us everything was fine." The stock opened bad and got worse. Margin compression on AI system sales. Not a one-time thing. A structural problem.
The Rotation Was Already Happening
The Nasdaq fell 1.69% to 23,195.17. Not carnage. But directional. The S&P 500 dropped 1.07% to 6,827.41. The Dow, though? Up 0.21% to 48,806.81. Goldman Sachs up 5% for the week. Johnson & Johnson up more than 4%. UnitedHealth up 3%.
Value stocks were winning. Blue chips. Boring holdings. The kind of companies your parents own. The kind that pay dividends and don't require you to understand quantum computing.
People were pulling out of Nvidia. Out of the AI infrastructure complex. Out of AVGO, out of Oracle (down another 3% Friday after Thursday's 10% collapse). Out of the Nasdaq, which had just been so close to new record highs. Friday ended it. The moment passed. You don't get another one like that this year.
Bitcoin Watches and Waits
Bitcoin dropped below $90,000 for the first time in weeks. Not a crash. A signal. It closed Friday around $92,000, but the intraday low of $89,800 was a message: when equities sell tech, crypto feels it immediately. The relationship is still there. Risk-on, risk-off. Bitcoin has spent the past seven days pinned between $88,000 and $94,000, which is precisely where you want to be when you have no conviction about what comes next.
Ethereum traded narrowly around $3,200, slipping 0.77%. Layer 1 assets rose 1.33% (Zcash up nearly 18%?), meme tokens up 1.45%, but the headline story is stillness. The crypto market is watching.
Altcoins were getting destroyed. JUP, KAS, QNT all posted double-digit losses on the week. CoinMarketCap's altcoin season index fell to a cycle low of 16/100. The memecoin index is down 59% year-to-date versus the CD10 down just 7.3%. The retail money is getting eviscerated. Only the institutional players—the ones that bought at the lows—are still positioned.
Who Actually Made Money This Week?
Lululemon. LULU jumped 10.7% Friday. CEO leaving at the end of January, and somehow the stock went up. The market heard "new leadership" and said "finally." That's your tell: markets don't care about continuity anymore. They want change. They want to see management teams that aren't drinking their own Kool-Aid about markets that never materialized the way the PowerPoint said they would.
Wealthfront (WLTH) opened at its IPO price on the Nasdaq Friday, which means it went sideways. A robo-advisor IPO in December 2025. The company has $88.2 billion in assets on platform, 1.3 million customers, 359 employees. That's a 245-to-1 customer-to-employee ratio. That should be the deal of the century. Instead it opened flat. The market yawned.
Quanex Building Products jumped 9% on earnings. Airlines got upgraded: UBS put a buy on American Airlines with a $20 price target (34% upside). Deutsche Bank upgraded Allegiant Travel to buy. Suddenly everyone's excited about flying. About hotels. About things that used to be called "value plays."
The Real Number Nobody's Talking About
Initial jobless claims spiked 44,000 the week before, hitting 236,000 total. That number came out December 12. The market barely flinched. But it's the real story. The Fed cut rates because it thought the labor market was faltering. Well, it is. Claims went from 192,000 to 236,000 in one week. That's velocity in the wrong direction.
Chicago Fed President Goolsbee—who voted against the December cut—said Friday he expects more cuts in 2026 than the median projection. He's being polite about it, but he's also saying the Fed made a mistake this week. Or will make one. Or both.
What Actually Happened Here
You trusted that the AI trade would paper over every other problem. You trusted that marginal utility would keep pushing, that enough money chasing GPU capacity would justify infinite capex. You trusted Oracle and Broadcom and Nvidia because the growth story was so big that profitability seemed like a problem for later.
Friday, the market decided later is here.
Goldman Sachs forecasts 12% S&P 500 earnings growth in 2026. Solid. Boring. The mega-cap tech stocks contribute 46% of that growth but only account for 26% of earnings. That's a concentration problem. If they stumble—if margins compress, if capex doesn't pay off, if AI ROI stays negative for longer than expected—the entire 2026 thesis breaks.
The Dow will be fine. Value will be fine. Banks will be fine. Tech will be fine if Nvidia actually delivers on inference and doesn't have its own Broadcom moment.
But the thing that went up the most all year? It's coming down. And nobody's caught it yet. Friday was just the opening.
What's Priced In Now
Markets expect two rate cuts in 2026. Maybe three if things really deteriorate. The Fed is done for 2025 after three cuts that didn't save the AI capex cycle. Bitcoin is waiting. The Dow is celebrating. Tech is wounded but not broken.
And somewhere on Wall Street, someone just canceled the $50 million Texas data center expansion because the ROI model doesn't work anymore without another 100 basis points of rate cuts.
The Fed can't give you 100 basis points. Maybe 50. Maybe nothing. The AI party just got a lot smaller invitation list.