The bond market just told equity holders exactly what their AI positions are worth in a world where the thirty-year yield clears five percent, and the answer emptied the Nasdaq of $1.3 trillion in two sessions. Friday's 4.2-percent drop—the index's worst day since April 2025—didn't come from a headline or a hack or a single disappointing print. It came from arithmetic: growth stocks priced for cuts got repriced for reality when the long end of the curve broke a level it hasn't seen since 2007.
The trigger was prosaic. The economy added 172,000 jobs in May, far exceeding forecasts, and the 10-year note finished June 5 at 4.55% while the 2-year note ended at 4.17%, its highest level since February 2025. The thirty-year yield topped 5%, fueling concerns that elevated borrowing costs could weigh on economic growth and investment. That five-percent handle is the line—the point where duration becomes a liability instead of a lever, where the present value of far-off cash flows collapses, and where Nasdaq darlings trading at fifty times forward earnings become fifty times too expensive.
The damage was surgical. Broadcom fell more than 7% following a double-digit decline on Thursday, while Marvell Technology and Micron Technology plunged about 16% and 13%, respectively. The S&P 500 lost 2.6% and the Dow Jones declined 1.4% after reaching a record high the previous day. This wasn't sector rotation; it was triage—selling the longest-duration, highest-multiple names first and asking questions later. The market that spent nine consecutive weeks ignoring macro risk because AI would save everything just discovered that AI doesn't save you from the cost of capital.
Here's what nobody's saying: this isn't about one jobs print or one Friday. A majority of Fed officials at the April meeting highlighted that some policy firming would likely become appropriate if inflation were to continue to run persistently above 2 percent, and prediction markets currently assign a 99% chance to no change at the June 16-17 meeting—but markets now price a 51% probability of a Fed rate hike in 2026. The gap between "holding steady" and "hiking again" is where positioning dies. Equity longs spent the last quarter pricing in eventual cuts; bond traders spent it pricing in the opposite. Friday was the day equity had to reconcile.
The tells were everywhere if you knew where to look. Polymarket puts 57% odds on zero cuts across all of 2026 against a Fed dot plot that still calls for one. The curve's been screaming this for weeks—you don't push the two-year to sixteen-month highs in a cutting cycle. You do it when the market realizes the cutting cycle isn't coming, that oil above $90 and payrolls above consensus and Warsh in the chair means the Fed has no room to ease, and that the entirety of Q2's equity gains were built on a premise that just evaporated. Growth rallied on the promise of cheaper money; that promise is now getting marked to a market where the thirty-year pays five and the Fed's next move might be up, not down.
- June 16-17 FOMC decision: Warsh's first meeting as chair. A hold is priced at 99%, but the statement language on future moves will be the real tell—any hint that hikes aren't off the table and you get another leg down in duration.
- The five-percent line on the thirty-year: If it holds above that level into next week, the mechanical selling from risk-parity funds and vol-control strategies hasn't finished. Friday was discretionary; Monday could be systematic.
- Semiconductor follow-through: Chips led both days of the break. If Nvidia and Broadcom can't find a bid with the VIX only in the low twenties, the nine-week win streak that just ended wasn't a correction—it was the top.