|
A Memory Chip Company Just Crossed $1 Trillion. The Bond Market Didn’t Flinch.
At 2:47 p.m. Eastern on Tuesday, Micron Technology crossed $1 trillion in market capitalization. The trading floor at the NYSE barely paused. A year ago the company was worth $87 billion. One UBS note tripled the price target from $535 to $1,625, and 19% happened in a single session.
On the same afternoon, Northland Capital downgraded Intel, warning that hyperscaler spending could decline by 2027. Two signals, same hour, opposite conclusions about the same AI buildout. That tension is the story the headline missed.
The spread that matters. The S&P 500’s forward earnings yield sits at 4.78%, per StreetStats. The 10-year Treasury closed Tuesday at 4.5%. A six-month T-bill pays 4.3%. The gap between what stocks offer and what bonds guarantee is 28 basis points, a quarter of a percentage point. That is the entire compensation the market is offering for bearing equity risk. The last time the premium was this thin was November 2021. The S&P peaked at 4,797 on January 3, 2022, and fell 25% over the next ten months. Nobody rang a bell. But the arithmetic was the same. When compensation for volatility shrinks to near zero, the market does not need bad news to correct. It needs the absence of good news.
Where the earnings have to come from. Consensus expects S&P 500 earnings to grow 21% this year. The 2027 estimate is 14%. The 2028 estimate is 11%. That three-year compound rate of roughly 16% is doing all the work. It is the only reason a forward P/E above 21 does not immediately repel institutional capital. If those estimates hold, the math survives even with the 10-year at 4.5%. If they slip by two or three percentage points, the case for equities over Treasurys collapses at current prices.
Micron is Exhibit A for the optimistic case. UBS projects more than $400 billion in cumulative free cash flow between 2027 and 2029. HBM supply is sold out through 2026 under long-term agreements. That is the kind of number that justifies a trillion-dollar valuation, but only if AI spending accelerates on schedule. Northland’s Intel downgrade landed the same afternoon, warning that hyperscaler budgets could contract by 2027. Two signals in the same hour, pointing in opposite directions about the same AI buildout. The base case here is that both are partly right, and the market is pricing only one of them.
The Dow told a different story. While the Nasdaq and Russell 2000 surged to records, the Dow fell 118 points. The 30-stock index is heavier in industrials, financials, and consumer staples. When tech-heavy indexes set records and the Dow declines on the same session, the rally is narrowing, not broadening. The last time that pattern appeared with this frequency was November 19, 2021. The S&P peaked six weeks later. Meanwhile, traders are pricing an 80% probability of a Fed rate hike by December, per CME FedWatch. That is a complete reversal from early 2026, when two cuts were expected. The Iran-driven oil shock pushed headline CPI to its highest level in nearly three years. Core PCE in February printed at 3.0%, well above the Fed’s 2% target.
Wednesday’s number. The BEA releases April PCE data on May 28. If core holds near 3% or ticks higher, the rate-hike probability climbs and the 10-year pushes toward 4.7%. At that level, the equity risk premium turns negative for the first time since 2007. For a reader holding a standard 60/40 allocation, that arithmetic matters. The bond side would carry better expected returns than the equity side at current prices. A cooler print near 2.7% would give the rally oxygen. But even then, the premium stays thin. My read: the question is no longer whether stocks can keep climbing, but whether they are paying you enough to justify being in them.
Two signals hit the tape in the same hour on Tuesday. One added $150 billion to a single stock. The other quietly noted that AI spending may peak sooner than anyone modeled. The market celebrated the first and ignored the second. That is usually how the arithmetic starts to turn.
|
