|
The Rally Was Built on Rate Cuts. The Jobs Report Took Them Away.
At 8:30 a.m. Eastern on Friday, the BLS reported 172,000 nonfarm payrolls added in May. The consensus had expected 80,000 to 105,000. March was revised up to 214,000. April was revised up to 179,000. Combined revisions added 93,000 jobs that the market had not priced in. By the close, the Nasdaq had lost more than a thousand points.
The catalyst was not the number itself. It was what the number removed from the table.
What the nine-week rally was pricing. From early April through Thursday’s close, the S&P 500 gained roughly 12%. The thesis was threefold. First, AI earnings were real: Dell, Snowflake, Micron, and HPE all delivered historic quarters. Second, the Iran ceasefire held long enough to pull oil from $126 to $88, easing inflation pressure. Third, and most importantly, the market assumed that softer economic data would give the Fed room to cut rates in the second half of 2026. That third pillar was doing more work than the other two combined. When 172,000 jobs landed on Friday morning, that pillar collapsed.
Why strong jobs are now bearish. PCE inflation sits at 3.8%. The fed funds rate is at 3.50–3.75%. The gap between the two is close to zero. In that configuration, a strong labor market means the Fed has no cover to cut. If anything, 172,000 jobs with rising wages gives the new chair reason to hike. CME FedWatch now shows hike expectations climbing heading into the June 16 FOMC meeting. The 30-year yield crossing 5% on Friday tells the bond market’s verdict before Warsh even speaks. The cost of capital just went up, and every stock priced for a low-rate world repriced lower in a single afternoon.
Why chips took the worst of it. Semiconductors led the sell-off for a specific reason beyond the macro. AI chipmakers trade at the longest duration in the S&P 500. Their valuations depend on earnings projected three to five years out, discounted back to the present. When yields rise, those distant cash flows are worth less today. A 50-basis-point move in the 10-year reduces the present value of 2029 earnings by roughly 8–10%. That is simple duration math. Applied to a sector that had run 40–60% year-to-date, the result was Friday’s rout. Marvell down 16%. Micron down 13%. AMD and Intel down 11%.
Broadcom’s Thursday sell-off set the tone. The market was already nervous after Hock Tan held the $100 billion full-year AI target. Friday’s macro trigger arrived into a sector that was primed to sell. A hawkish jobs print plus an AI sentiment crack produced the worst two-day chip rout since the tariff shock of April 2025.
What happens next. The SpaceX IPO prices on June 11 and lists on June 12. Warsh chairs his first FOMC on June 16–17. Both arrive into a market that just lost its rate-cut thesis. The chip sector is in correction and 30-year yields sit above 5%. The base case here is that Friday was not a one-day event. It was the repricing that begins when the economy proves too strong for the monetary policy the market was expecting. The nine-week rally was earned by real earnings. The valuation it produced depended on rate cuts that are no longer coming.
My read: the AI buildout is not over. The earnings were real. Dell’s $16 billion AI quarter was real. Snowflake’s $6 billion AWS deal was real. But the stocks were priced for those earnings at a 4% 10-year, not a 4.5% 10-year with the 30-year above 5%. Friday was the day the discount rate changed. Everything downstream from that changes with it.
The economy added 172,000 jobs. The Nasdaq lost 1,121 points. The jobs were real, but the rate cuts were not. The market spent nine weeks pricing one and ignoring the other. On Friday, it stopped ignoring.
|