The consensus reading of Wednesday's global risk-on move is that a geopolitical relief headline produced a melt-up in an already expensive market, and Michael Burry's warning of a 1987 echo following the S&P 500's new closing high has become the fashionable frame for that view. It is the wrong frame, because 1987 was a multiple event and this is not one. The index had gone 42 sessions without a closing high while companies in it were tracking earnings-per-share growth of nearly 50% for the spring quarter, according to FactSet — which means the market spent six weeks de-rating against accelerating profits, and the record simply marks the point where the arithmetic caught up.

a computer screen displaying a stock market chart
Photo by lonely blue via Unsplash

Where We Are in the Cycle

A closed or contested Strait of Hormuz is not primarily an earnings event for the largest companies in the index; it is a discount-rate tax, expressed through energy prices, through the inflation path they imply, and through the term premium investors demand for holding long-duration cash flows against that uncertainty. Optimism that oil traffic could resume, alongside a U.S.–Iran memorandum of understanding, is therefore being transmitted less through the profit line than through the denominator, and the pattern across Asia — the Nikkei up 3.0% and South Korea up 3.4%, with oil far below recent highs and bonds firm — is exactly what an unwinding risk premium looks like when it arrives all at once. Cycle-wise, this is mid-cycle earnings acceleration being un-discounted rather than late-cycle multiple expansion being chased, and the two regimes carry opposite implications for how much of the move is durable. The distinction matters because a de-rating that reverses can run considerably further than a re-rating that extends.

Session snapshot · Aug 4, 2026
SymbolCloseChangeDay range52-week range
AMD$525.00+40.36 (+8.33%)$502.41–$530.06$149.22–$584.73
V$369.32+3.65 (+1.00%)$361.00–$371.08$293.89–$373.97
JPM$357.55+4.91 (+1.39%)$354.62–$363.00$279.10–$363.00
TSLA$327.26+5.18 (+1.61%)$320.80–$329.53$297.38–$498.83
META$587.75-2.49 (-0.42%)$578.26–$592.73$520.26–$796.25
AMZN$277.33-6.69 (-2.36%)$275.86–$281.04$196.00–$287.16
GOOGL$377.61+4.10 (+1.10%)$367.50–$380.52$190.12–$408.61
NVDA$211.90+5.26 (+2.55%)$209.05–$213.06$164.07–$236.54
MSFT$492.65+5.00 (+1.03%)$479.39–$499.34$349.20–$553.72
AAPL$309.34+5.91 (+1.95%)$301.38–$310.42$201.68–$344.57

What the Session Revealed

AMD was the headline, closing at $525.00 for a gain of 8.33% after gapping 4.04% higher and finishing in the upper third of its range — outperforming NVDA by 578 basis points on a day when both carried the same end-market narrative. That leadership then developed a crack after the bell, with AMD's post-earnings pullback keeping Nasdaq futures flat, which is the essential lesson of the semiconductor complex at this stage: it is priced richly enough that individual prints, not macro relief, now set its path. AMD remains 10.2% below its 52-week high of $584.73, and NVDA sits 10.4% below its own at $236.54, so neither name is confirming a breakout regime despite the size of the move.

us a flag on top of building
Photo by David Vives via Unsplash

The more informative behavior sat in the names nobody was talking about. JPMorgan traded to $363.00 intraday — precisely its 52-week high — before closing at $357.55, and Visa finished at $369.32, just 1.24% below its own 52-week high, while Meta closed at $587.75, fully 26.2% beneath its high of $796.25. When the financial complex is pressing annual highs and the most crowded advertising-AI asset is a quarter below its peak, the leadership of this advance is not the leadership the narrative assumes.

Six of the ten largest names opened below their prior close, including Microsoft, Alphabet and Visa — and every one of the six finished above its opening print, with Microsoft recovering 2.47% and Alphabet 2.55% from the open.

Amazon was the exception that proves the mechanism: it gapped 2.7% lower, could not repair, and closed at $277.33 in the lower third of its range, 3.4% below the 52-week high it had touched days earlier. My call that Amazon would clear $287.16 is now the weakest of the pending set, while the AMD above $490.95 call cleared with room to spare.

The Positioning Implication

The consensus expression of a Hormuz reopening is straightforward and largely spoken for: sell energy, buy transports and consumer discretionary, and add to the AI complex because every risk-on day flows there by reflex. That last leg is the crowded one, and AMD's after-hours reversal is the tell — a position whose valuation requires each quarter to be a beat is a position sized against event risk, not against a cycle. The less appreciated beneficiary is the balance-sheet-heavy financial complex, which never received an AI premium in its multiple, which gains directly from an easing rate path and a compressing risk premium, and which is nonetheless the part of this group closest to its annual highs. Investors would do well to consider that Tesla, at $327.26, sits only 10.05% above its 52-week low of $297.38 and 34.4% below its high — cheaper energy is a headwind to that franchise, and its participation Wednesday was the weakest form of confirmation on offer.

a person holding a tablet with a chart on it
Photo by Coinstash Australia via Unsplash

The Weight of Evidence

The record high is earnings arithmetic asserting itself after six weeks of geopolitical de-rating, and the argument lives or dies at JPMorgan $363.00: a close above that level within ten sessions confirms the discount-rate interpretation and puts cyclical breadth behind the index, while semiconductors give back some of Wednesday's gain — AMD below its session low of $502.41 within three sessions would be consistent with the thesis rather than damaging to it. What would reverse my view is the mirror image: financials failing back through $354.62 while chips carry the index alone, because a rally that narrows into the most expensive cohort while the cyclical complex rolls over is the 1987 analogy earning its keep. The evidence today does not support that, and I am positioned for the de-rating to keep unwinding into cyclicals.