Replay drill: the AI-tech rotation week of June 22-26, 2026

Data cutoff: U.S. cash-market close on Friday, June 26, 2026, at 4:00 p.m. ET. This replay is educational only. It is not a prediction, a recommendation, or a model portfolio. The strategy test below ignores commissions, bid-ask spread, taxes, slippage, position sizing, and the emotional pressure of real money.

The week in rewind

This was a useful practice week because the headline index did not tell the whole story. The S&P 500 ETF, SPY, fell 2.38% from the prior close to Friday's close, while QQQ fell 4.60% and NVDA fell 8.62%. The Dow, by contrast, finished the week higher in the broader index recap published by T. Rowe Price. That is the kind of split tape where retail traders often feel as if the market is sending two messages at once. 1 2 3 4

Day	Opening setup	What hit	What actually moved
Monday, June 22	Traders came in after a tense weekend around U.S.-Iran peace talks, with mega-cap tech already under pressure.	Big tech selling weighed on the Nasdaq, while the Dow managed a small gain.	SPY closed down 0.31%; QQQ closed down 0.36%; NVDA closed down 0.97%. 5 1
Tuesday, June 23	The first real test was whether Monday's tech weakness would stay contained.	The selloff widened into memory and chip stocks, and SpaceX rebounded from a record low.	SPY fell 1.45%; QQQ fell 3.29%; NVDA fell 4.13%. 6 2
Wednesday, June 24	Dip-buyers had a cleaner setup: oil was falling, but tech was still being sold.	The S&P 500 and Nasdaq fell for a third straight day while oil dropped to its lowest level since the start of the war.	SPY was nearly flat on the close, down 0.05%, but QQQ still lost 0.42%. 7 1
Thursday, June 25	The macro tape moved back to inflation.	BEA reported May PCE inflation at 4.1% year over year and core PCE at 3.4%, with income and spending both up 0.7% for the month.	SPY rose 0.14%, but QQQ gained only 0.81% and closed well below its opening price. 8 9
Friday, June 26	The question was whether the AI trade could stabilize before the weekend.	Reuters reported a 5.3% drop in the PHLX chip index and a 7.9% weekly loss for that chip index, while CNBC reported rotation into health care, staples, financials, and utilities.	SPY slipped 0.72%; QQQ fell 1.38%; NVDA fell another 1.64%. 10 11

Read the week chronologically and the lesson changes. This was not simply "stocks were down." It was a rotation week. The most crowded winners of the first half were hit hardest, while defensive groups and the Dow held up better.

Decision points

Freeze 1: Monday at the open, your portfolio is tech-heavy

What you knew then: tech had been weak, U.S.-Iran headlines were still in the background, and the market had not yet shown whether selling would spread beyond a few crowded names. You did not know Tuesday would be the heavy down day.

Three reasonable choices:

Trim a small slice of tech exposure before the pattern confirms.
Hold everything because one soft day is not a trend.
Rotate a portion into broader or defensive exposure, not because you know tech will fall, but because concentration risk has risen.

Reveal: the rotation did spread. QQQ lost 4.60% for the week and NVDA lost 8.62%, while the Dow outperformed. Choice 1 and choice 3 would have reduced pain; choice 2 was not automatically a bad decision if it matched a written plan. The bad version of choice 2 is "I froze and called it conviction afterward." 3 4

Freeze 2: Tuesday close, SPY has dropped 1.45%

What you knew then: SPY had a down day big enough to feel like a classic "buy the dip" signal. You also knew the selling was not random; it was hitting tech and chips in particular.

Three reasonable choices:

Buy SPY at the next open for a one-day mean-reversion trade.
Wait for the close to see whether the next session confirms support.
Reduce single-stock chip exposure but keep broad index exposure intact.

Reveal: the simple one-day dip trade lost 0.26% from Wednesday's open to Wednesday's close. That is a small loss, but it matters because the trade did what it was supposed to do and still failed. A clean setup can lose. A messy setup can win. Hindsight bias tries to label the outcome as proof of decision quality; your journal should not let it. 1

Freeze 3: Thursday at 8:30 a.m. ET, PCE is hot but spending is firm

What you knew then: BEA had just reported headline PCE inflation at 4.1% year over year, core PCE at 3.4%, and 0.7% monthly gains in both personal income and spending. That is not a clean bearish or bullish print. It supports the consumer story and keeps the inflation problem alive. 8

Three reasonable choices:

Sell first and ask questions later because higher inflation raises rate-risk.
Do nothing until the market shows which part of the report it cares about.
Stay invested but shift attention from AI beta to sectors showing relative strength.

Reveal: the broad market did not collapse on Friday, but chips did. Reuters reported the PHLX chip index down 5.3% on Friday, while CNBC noted strength in health care and defensive groups. Choice 3 would have matched the tape best. Choice 2 was defensible if your rule is to avoid trading the first interpretation of a data print. Choice 1 was emotionally understandable, but it risked treating a mixed report as a single-message event. 10 11

Strategy lab: one-day buy-the-dip on SPY

Rule tested: when SPY closes at least 1.0% below the prior close, buy SPY at the next session's open and exit at that same session's close. Hold cash otherwise. No fees, slippage, taxes, or position sizing adjustments are included.

This week produced one signal: Tuesday's 1.45% SPY drop. The rule bought Wednesday's open at 735.17 and exited Wednesday's close at 733.24, for a one-day loss of 0.26%. 1

For the longer window from January 2 through June 26, 2026, the same rule produced 15 trades, won 7 of them, averaged +0.09% per trade, and compounded to +1.24% before costs. That is not useless, but it is fragile. A few cents of spread, a late fill, or one emotional override can erase a system whose average trade is less than one-tenth of one percent. 1

Known failure modes:

It buys weakness without asking whether the weakness is broad-market fear or a sector-specific unwind.
It exits after one day, so it can miss slower recoveries.
It looks smoother in a spreadsheet than it feels when the opening print gaps against you.
Its edge, if any, is too small to treat as permission to oversize.

Your trade journal

Use these prompts before you look at next week's watchlist:

Which trade from last week would you repeat even though it lost money?
Which winner came from a weak decision that happened to work?
Where did you confuse "I have a plan" with "I do not want to act"?
If you were forced to reduce one source of concentration risk before Monday's open, what would it be, and why?

Write the answers in plain language. If the answer sounds like a press release, keep going until it sounds like something you would say to another trader at your desk.

Lesson of the week

A good replay does not ask, "Could I have predicted Friday?" It asks, "At each frozen moment, did my action fit the information I actually had?"

Replay drill: the AI-tech rotation week of June 22-26, 2026

The week in rewind

Decision points

Freeze 1: Monday at the open, your portfolio is tech-heavy

Freeze 2: Tuesday close, SPY has dropped 1.45%

Freeze 3: Thursday at 8:30 a.m. ET, PCE is hot but spending is firm

Strategy lab: one-day buy-the-dip on SPY

Your trade journal

Lesson of the week

참고 출처

관련 콘텐츠