H2 Opens With the AI Trade Losing Its Monopoly

The tape: a tale of two indices

The first two trading days of H2 2026 produced one of the sharpest leadership reversals of the year. On July 2, the Dow Jones Industrial Average closed at 52,900.07, up 594.83 points (+1.14%), while the NASDAQ Composite fell 207 points to 25,832.67 (-0.80%).{{cite:chatcmpltool}} The S&P 500 finished essentially flat at 7,483.24 — but that single print masks a violent tug-of-war beneath the surface.

The semiconductor ETF (SMH) peaked at $655.89 on June 30 — the last session of the first half — then plunged to $592.29 by the July 2 close, a two-session decline of 9.7%.{{cite:chatcmpltool}} NVDA fell from $200.09 to $194.83 over the same window, down 2.6%.{{cite:chatcmpltool}} Meanwhile, the Health Care Select Sector SPDR (XLV) rose from $158.66 to $163.74, up 3.2%, as capital rotated from the AI complex into defensives.{{cite:chatcmpltool}}

Analysts dubbed the move a “Great Rotation”: chip stocks saw profit-taking after an 80% first-half surge that added roughly $2 trillion in combined market capitalization to the semiconductor complex, with capital shifting toward blue chips and software names.{{cite:chatcmpltool}} The July 1 selloff alone erased more than $200 billion in combined market value from Micron, Intel, and AMD.{{cite:chatcmpltool}}

Three catalysts converging

The rotation was not driven by a single trigger. Three forces hit simultaneously, each reinforcing the others.

1. Profit-taking after a historic H1 run

Semiconductor stocks surged roughly 80% in the first half of 2026 on AI infrastructure demand.{{cite:chatcmpltool}} After a run of that magnitude, the mere turn of the calendar to Q3 was enough to prompt de-risking. As one analyst framed it, H2 opened with “the AI trade losing its monopoly” over market leadership — the S&P 500 still near highs, the Dow making fresh ones, but beneath the index level, “a fairly serious change in leadership.”{{cite:chatcmpltool}}

2. The AI memory glut scare

Micron shares fell over 9% despite strong Q3 earnings, as a DRAM lawsuit and fears of a future supply glut overshadowed booming AI-driven memory demand.{{cite:chatcmpltool}} The contradictions are stark: Micron’s HBM chips are sold out through 2026 and DRAM prices have surged up to 6x, yet rising capex across the industry raises the specter of overcapacity.{{cite:chatcmpltool}} The question the market is suddenly asking is whether the AI capex cycle is building revenue — or just building capacity that will eventually compete against itself.

3. A weak June jobs report

The U.S. economy added just 57,000 jobs in June, roughly half the 100,000 forecast, as 507,000 workers left the labor force — making the dip in the unemployment rate to 4.2% deeply misleading.{{cite:chatcmpltool}} Wage growth tracked below inflation for a third consecutive month.{{cite:chatcmpltool}} For a market already questioning whether AI spending will pay off, a cooling labor market adds a second leg of uncertainty: if employment is softening while the Fed holds rates restrictive, the growth cushion narrows.

Apple: the single stock holding the line

While chip stocks cratered, Apple surged nearly 5% on July 2, adding roughly $182 billion in market capitalization in a single session and single-handedly preventing what would have been a much sharper S&P 500 decline.{{cite:chatcmpltool}} The catalyst: Apple told suppliers to prepare to make approximately 10 million foldable iPhones this year, up from a previous forecast of 7–8 million units.{{cite:chatcmpltool}}

AAPL closed at $308.63 on July 2, up from $281.74 on June 29 — a 9.6% four-session rally.{{cite:chatcmpltool}} The irony is worth noting: the same market questioning whether AI infrastructure spending will generate returns is rewarding a consumer hardware play with tangible unit economics. The rotation is, in part, a flight from capex promise to product proof.

The Fed backdrop: Warsh draws the line

Federal Reserve Chairman Kevin Warsh, speaking July 1 at the ECB Forum on Central Banking in Sintra, Portugal, vowed to “disappoint” anyone who expects loose monetary policy, reaffirming his commitment to the 2% inflation target.{{cite:chatcmpltool}} He declined to signal what the central bank may do at its July meeting but noted that inflation was “too elevated.”{{cite:chatcmpltool}}

The macro dashboard explains his posture. The FRED snapshot puts CPI inflation at 4.17% year-over-year — more than double the target.{{cite:chatcmpltool}} The Fed funds rate sits at 3.63%, having been cut by 70 basis points over the past year, yet real rates remain restrictive given the inflation overshoot.{{cite:chatcmpltool}} UBS pushes back on market pricing for further Fed tightening, arguing that cooling wages and fading tariff effects favor a hold.{{cite:chatcmpltool}}

Consumer sentiment, which collapsed to a record low of 44.8 in May during the U.S.-Iran conflict, recovered to 49.5 in June as gasoline prices moderated — but remains 18.5% below its year-ago level of 60.7.{{cite:chatcmpltool}} Real GDP growth of 2.66% YoY and a 4.2% unemployment rate suggest an economy that is slowing, not collapsing.{{cite:chatcmpltool}} The yield curve (10-2Y) is positively sloped at +31 basis points, and high-yield credit spreads remain tight at 2.75% — neither flashing imminent distress.{{cite:chatcmpltool}}

The historical analog: 2006, or 2007?

The FRED macro snapshot identifies mid-2006 and October 2007 as the closest historical analogs to the current environment, with similarity scores of 0.95.{{cite:chatcmpltool}} Both periods featured unemployment near 4.6–4.7%, CPI inflation in the 3.6–4.1% range, and a Fed that had tightened and was holding firm.

The analogy cuts both ways. Mid-2006 was a mid-cycle slowdown that resolved without recession — the Fed paused, growth reaccelerated, and the market continued higher into 2007. October 2007, however, was the last calm before the storm: the macro snapshot looked benign, but the financial system was already fracturing beneath the surface.

What separates the two outcomes is not the macro print but the hidden leverage. In 2007, subprime mortgages were the concealed risk. In 2026, the question is whether AI infrastructure capex is the analogous concealed risk — a cycle where spending has outraced revenue, and where a glut scare in memory chips could be the first sign that supply is outpacing demand.

What would have to be true

Two interpretations of the rotation are competing, and each requires specific conditions to hold.

The healthy read — breadth expansion, not breakdown:

• Profit-taking after an 80% H1 surge is normal, not pathological.
• Capital broadening from a narrow AI trade into defensives and blue chips is what a late-cycle market needs to sustain a bull run.
• Q2 earnings season shows AI capex translating into revenue, not just capacity.
• The Fed holds steady, real rates stabilize, and the labor market cools without cracking.
• If these hold, the rotation is a healthy breath — the kind that extends bull markets.

The warning read — the first crack:

• A two-session 10% semiconductor decline triggered by a glut scare, against a backdrop of a weak jobs report and a Fed chair vowing to hold the line at 4%+ inflation, has the shape of a regime change.
• The memory oversupply fears prove real — capex has built capacity faster than demand can absorb it.
• The labor market cooling accelerates beyond June’s 57,000 print.
• Warsh’s hawkish vow keeps real rates restrictive even as growth slows, tightening the financial conditions vise.
• If these hold, this is not a rotation — it is the first crack.

The base rate favors the healthy read: most sharp sector rotations after a strong half resolve into breadth expansion rather than systemic breakdown. But the October 2007 analog — 0.95 similarity, macro looking fine, storm underneath — is the reason the warning read cannot be dismissed.

What to watch next

• Q2 earnings season — Does AI capex translate into revenue, or just capacity? Micron’s glut scare will be tested by the chipmakers’ own forward guidance.
• July FOMC meeting — Warsh declined to signal a move. A hold is the base case, but the market is pricing tightening odds that UBS calls overdone. The statement and dot plot will matter.
• Labor market trajectory — June’s 57,000 payrolls and 507,000 labor-force exits bear watching. A second soft report would shift the narrative from “cooling” to “deteriorating.”
• Consumer sentiment — June’s recovery to 49.5 is fragile. If gas prices re-accelerate or geopolitical tensions resurface, sentiment could retest the May record low of 44.8.
• Semiconductor breadth — Whether the SMH’s 9.7% two-day decline stabilizes or extends will tell whether this is profit-taking or de-risking. Watch NVDA, Micron, and AMD for follow-through.