Market Wrap: The Week the AI Trade Hit Its First Real Wall -- June 2 to 6, 2026

My friend sent me a message on Friday night.

He said: what just happened to my portfolio?

He had been holding the same AI semiconductor stocks all year. He watched them climb 40, 50, 90 percent. He felt smart. He felt early. He felt like he understood something most people did not.

Then this week arrived.

I looked at my own portfolio. It took a hit too.

Market Wrap: The Week the AI Trade Hit Its First Real Wall -- June 2 to 6, 2026

It started so well.

June 1. The ISM Manufacturing PMI came in at 54.0. Highest since May 2022. Fifth straight month of expansion. The physical economy was accelerating at its fastest pace in four years. Every data point confirmed the story. The AI buildout is real. The money is flowing. The demand is not slowing.

That same night, two AI infrastructure companies reported earnings. Both beat analyst estimates. One dropped 12 percent. One surged 29 percent.

$Credo Technology (CRDO.US)$ had run 35 percent in the month before reporting. When the good news arrived, the crowd that had already bought the good news had nobody left to sell to. Stock fell.

$Hewlett Packard Enterprise (HPE.US)$ came in under-owned. Nobody expected what it delivered. Buyers rushed in from every direction. Stock surged 29 percent.

Same night. Same sector. Same AI infrastructure theme. 41 point swing between two companies. Not because one business was better than the other. Because of where the crowd was standing before the numbers landed.

June 2. $Palo Alto Networks (PANW.US)$ beat every number. Beat on earnings. Beat on revenue. Raised guidance. Stock spiked 10 percent in the first minutes. Then the crowd that had run it up 55 percent in a month used the spike as their exit. By Thursday it was trading below where it closed before the report.

June 3. The biggest night of the week.

$Broadcom (AVGO.US)$ reported its best quarter in company history.

Revenue of $22.2 billion. Up 48 percent year over year. AI chip revenue of $10.8 billion. Up 143 percent. Free cash flow crossed $10 billion in a single quarter for the first time in the company's history. EBITDA margin of 69 percent. A record on every single line.

The stock fell 13 percent.

Here is why.

There is a number that never gets written down. It never appears in any analyst report. It does not show up on any financial website. But every institutional trader knows it exists.

It is called the whisper number.

The official EPS consensus was $2.40. The whisper was $2.45. Broadcom delivered $2.44. One cent below the whisper.

The official Q3 AI chip guidance was $10.7 billion. The analyst whisper had moved to $17.2 billion. Broadcom guided $16 billion. A $1.2 billion miss against a number that was never written anywhere.

The crowd at $479 had priced in the whisper. When $16 billion arrived instead of $17.2 billion, there was nothing left to hold the price at $479. The crowd sold. Not because the business failed. Because the price had assumed something that did not come.

Then it spread.

June 4 and June 5. The logic moved through the entire semiconductor chain. If Broadcom, with 143 percent AI chip revenue growth, could not clear the whisper, what did that say about AMD, Micron, Intel, ASML?

$Advanced Micro Devices (AMD.US)$ fell 12.6 percent over two days. $Micron Technology (MU.US)$ fell 17 percent. $Intel (INTC.US)$ fell 9 percent. South Korea's Kospi dropped 5.5 percent in a single session. Over one trillion dollars in market capitalisation erased from the global semiconductor complex.

Then on June 5, the jobs report landed.

172,000 jobs added in May. The forecast was 85,000. More than double. The unemployment rate held at 4.3 percent. Previous months revised upward by a combined 93,000 jobs.

A strong economy. Good news for most people. But for high-multiple technology stocks already under selling pressure from the semiconductor rout, it was the second punch of the week.

Strong jobs means the Fed has less reason to cut rates. Higher rates compress the present value of future earnings. The stocks that had already been falling now had a second force pushing in the same direction.

Two forces. One direction. Five days.

$SPDR S&P 500 ETF (SPY.US)$ closed the week at $737.55. Down 2.64 percent. It had printed above $7,600 for the first time in history on June 2. Four days later it was back below $740.

$Invesco QQQ Trust (QQQ.US)$ closed at $705.06. Down 4.18 percent for the week. Its worst week since April 2025. It hit a 52-week high of $748.65 on June 3. By Friday it had given back nearly six weeks of gains in four sessions.

$Dow Jones Industrial Average (.DJI.US)$ held relatively better at minus 1.35 percent. Its blend of healthcare, financials, and industrials provided a buffer. But make no mistake. This was a broad risk-off week. Everything sold.

$iShares Russell 2000 ETF (IWM.US)$ closed down 3.55 percent. Small caps did not catch a bid. They fell alongside tech. This was not an orderly rotation. It was a market-wide selloff driven by two simultaneous forces that nothing was positioned for.

Now back to my friend. And back to my own portfolio.

We both took hits this week. I am not going to pretend otherwise. Positions in the AI infrastructure space that looked strong on Monday looked very different by Friday.

But here is what I keep coming back to.

Broadcom's AI chip revenue grew 143 percent year over year. HPE posted a record AI server backlog. The ISM confirmed physical expansion at the fastest rate in four years. The 172,000 jobs report confirmed the economy paying for all of this infrastructure is still healthy.

None of that changed between Monday and Friday.

What changed is that the crowd had built expectations the businesses could not clear in a single week. A sector up 90 percent in five months prices in perfection. This week proved it could not clear that bar. The correction is real. The recovery timeline is uncertain. But the structural story beneath the businesses did not change.

The business beneath the stock did not change this week.

That is the only distinction that matters.

Correction note: An earlier version of this post stated that the Russell 2000 rose 1.45 percent on Friday while the Nasdaq fell. This was incorrect. IWM closed June 5 at $281.65, down 3.55 percent. Small caps sold alongside tech. This was a broad risk-off week, not a rotation story. A reader who trades IWM daily caught the error and I am grateful for the correction. The rotation narrative has been removed from this post. Everything else stands.

770K Views