Microsoft clears 514 on volume. Meta takes the rest
US Stocks · Options · News · Views
The tape was broad. The Dow did the lifting
Friday closed green across the four averages. SPX opened 7710, high 7752, low 7693, closed 7743, up 0.5%. Nasdaq closed 27069, up 0.5%. The Dow closed 51829, up 0.9%, and led. Russell closed 2838. IWM closed 282, basically unchanged. QQQ closed 745, up 0.5%. VIX closed 14.87, down 5%.
329 S&P names rose, 172 fell. Tech, industrials, staples, and financials did the work. SOXX closed 573, up 1.2%. This is not Thursday's tape, when the index sat still and META carried it. Friday was actually majority green. Treasuries did not get out of the way. Oil eased. Breadth was fine. The discount rate is still high.
The 10-year closed 5.18%. That is not a pullback
The 10-year yield traded as high as 5.23% and closed 5.18%, two basis points above Thursday's 5.16%. TLT opened 79.37, high 79.46, low 78.83, closed 79.32, almost unchanged. The intraday round trip looks more like month-end repositioning than a trend reversal.
Thursday's 7-year auction was soft and tagged a short-term high in yields. Friday did not give 5.16% back. The question is whether 5.2% holds or whether yields fall back under 5% in the next few sessions. If they sit above 5.2% for long, growth stocks will feel it. Two days at these levels is not enough to call that. Next week ISM, PCE, and payrolls reprice the vol.
Oil got a more optimistic read on US-Iran contact. WTI closed 92.41, down 2.3%. Brent closed 104 and is still above 100. Finer technical talks are not a signed deal. Cheaper oil takes a bit of pressure off the index. It does not close the risk.
Durables were flat. Core capital goods were the hard print
August durable-goods orders were basically unchanged, better than the expected decline. Transportation fell 0.6% and did the damage. Ex-transport orders rose 0.3%, short of the 0.6% expected. Nondefense capital goods excluding aircraft rose 1.6% versus 0.5% expected. July was revised to about 0.6%. That line is up about 14% year over year. It is the cleaner proxy for business equipment spending. Computers and equipment are still adding. The semiconductor bid did not shut itself off in the order book. Strong orders do not let the index ignore a 5.2% discount rate. They only say the AI capex leg has not broken in the data.
Michigan's September final sentiment printed 48.1, above the 47.6 forecast and the 47.8 prelim, but still below August's 51.7. One-year inflation expectations are 4.6%. Five-year are 3.4%. The one-year rate has already come up from 4.0% in August. Inflation expectations are not rolling over. Jobs are steady, equipment demand is still there, and inflation expectations are not coming down. The discount rate will not ease itself.
Microsoft cleared 514 on volume. The new shelf dies at 490
MSFT opened 499, high 519, low 497, closed 516, up 3.7%. Volume was about 38.2 million shares versus 16.7 million on Thursday. It closed through 514. The 519 high is a multi-month high. That shelf had been in the way for days. Thursday was still grinding under 498. Friday went through it on size, not a quiet leak.
The same day Microsoft shipped a new Copilot: Home, Code, and Autopilot. Chat, Office, coding, and a persistent agent sit in one door. Office is embedded in Copilot. Autopilot can keep running when nobody is watching. Most of it is still preview and early-access. The tape paid for consumer-grade agents people can actually use, not for a rewrite of this year's subscription revenue. Confidence moved first. Volume followed. Thursday's seat discounts were not heard as a profit collapse. Friday is the product door. Same direction: put agents where people will use them.
Above 514 the tape is a higher shelf. A break of 490 is what turns the uptrend back into a range. Weekly RSI divergence is still there. After choosing the upside break, a smash-first repair is harder. The more likely path is stretch, then pull back. 600 is the more neutral upper end of the mid-2027 forward multiple band. It is not a calendar. Nobody can date a move to 600 by year-end. Long-term holders stay. Anyone who sold and waited for a dip already has a right-side break. A pullback probably still lives in 490 to 514.
Meta closed 752. The next buy is 690 to 730
META opened 769, high 770, low 747, closed 752, down 3.3%. Thursday closed 778, up 4.5%, high 780, already overbought. Friday's volume was about 25.8 million versus 34.8 million on Thursday. That is a stall-and-give-back, not a crash. The September 21 volume candle traded mostly in 690 to 730, with a low near 680. If you are flat and waiting, watch that band. Do not treat 752 as the discount. Whether it reaches 690 is unknown. A collapse is not the base case. Thursday entered the 745-785 pressure band. Friday came off the top of it. The next cheaper print is not 752.
GOOG closed 341, up 0.6%. AMZN closed 250, basically unchanged. 244 is still intact. Both are still long-term holds waiting for their turn. A late start is not a broken fundamental. Size the book yourself. Do not hear the range as a reason to run.
TSMC is on the 2027 mid. SOXX still needs 577
TSM opened 452, high 455, low 449, closed 451, almost unchanged. After 429 the near-term overhead is thin. The 2027 forward mid is about 457. The upper band is about 593. Spot is sitting on the mid. Tagging the 479 prior high is a matter of time, not a claim that volatility is gone. 2027 capex consensus has started to drift higher, including the large clouds and xAI. If that revision sticks, it helps the top of the hardware chain: TSM, ASML, NVDA, AMD, AVGO. The number is not in yet. Treat it as a seedling, not a new table.
SOXX closed 573, up 1.2%, high 576. 577 is still a few points away. Through it, the structure looks toward prior highs. Fail it, and it is still a box. Daily RSI divergence is there. Even a 577 break has to repair. The script is a FOMO push, a technical fade, then higher lows. It is not a straight shot to the June high at 656.
QQQ closed 745. The June high is about 749. It is not through. Near-term supply is thin, so a grind-up is easy. That is not a new high. Hold what you own. Chasing the right side with extra leverage has no clean payoff math. Hedge-fund cash is already thin and the extra leverage is not on. You do not sell, they do not add, and the index can crawl. SPY closed 771. The August high at 779 is still overhead. Support underneath is strong. VIX compressed again. September has one more week of data. Downside needs a surprise. Without one, you do not sell and you do not lever up.
Three lines. Microsoft has to hold 514, or 490 is the tell. The 10-year either stops near 5.2% or falls back under 5%. SOXX either takes 577 or stays in the box. Meta's next buy is 690 to 730, not 752.