What moved and why

Stocks closed higher on Friday. The Dow ended a three-session losing streak [16][21]. Oil and yields both eased: Iran offered to reopen the Strait of Hormuz within 7 days and restart nuclear talks [50]. US crude fell nearly 8% on the week [34], and Saudi exports hit their highest level since the war began, despite the pipeline outage [37]. The 10-year ended the day little changed [30] after touching a 19-year high this week [45]. Microsoft led the Dow and closed at its highest level of the year [18][23]. Tesla fell because the Semi launch did not excite investors [26].

What I cannot see: the panel's 1-day data is stale. Every name reads +0.00%, breadth is 44 unchanged, and the S&P 5d reads -0.27%, while the pre-open preview said the S&P was heading for a weekly gain [52]. I cannot measure the size or breadth of Friday's rally. The panel shows dollar volume -21% vs the 5-day mean, which may be a data artifact. Today's materials do not explain INTC +17.3% or AMD +12.4% over 5 days.

Updating yesterday's call: my 9/24 read ("Brent >$106, favor upstream oil") is contradicted by [34][37][50]. My Energy buys have lost at every horizon (9/14 7d excess -5.80%, 9/18 3d -1.87%). I am dropping the upstream long. The rates story has not reversed. Yields paused. Gold fell on the week because the market expects a hawkish Fed [10]. Warsh's framework still leaves room for more hikes [56].

Sector reads

• Energy — neutral (was buy). Barrels now carry diplomatic downside: a 7-day reopening offer [50] plus record Saudi exports [37]. Trump is weighing ways to lower diesel costs without a full export ban [2]. That removes the tail risk behind my 9/22 refiner avoid. Refiners (VLO -7.36% 5d) now look better than upstream on a relative basis. I am not buying either yet.
• Utilities / Real Estate — avoid. A 7.45% mortgage rate, the highest since April 2024 [78], and solar names selling off on financing costs [82] tie directly to long rates. Yields paused on Friday; they did not fall [30]. XLU is -8.85% and XLRE -6.74% over 20d. This call is wrong if the 10-year closes below its pre-rout level.
• Communication Services — hold META, don't add. Connect hardware and Muse keep the story going [77][79][20]. Options pricing supports continuation [39]. Still, META is +36% in 20 days, and my buy calls average -1.39% excess at 20d. Keep the position, do not chase.
• Technology — selective; Micron is the event. MSFT is strong [18]. Micron reports next week with pricing power as memory supply stays tight [33][73]. One analyst warns revenue may shift into next year, depending on when Nvidia's new product ramps [25]. There are cracks on the capex side. Oracle sent a force majeure notice on its New Mexico data-center project [58]. Bernstein flags iPhone margin risk from component costs [24]. Keep the small MU long through the print. Stay cautious on data-center supply-chain names (VRT -8.91% 20d), though the materials do not link Oracle to VRT directly.

What matters next

1. Micron earnings (next week; the exact date isn't in today's materials) [33][73]. Guidance that pushes revenue into next year [25] would test the memory leg of the AI trade. MU is +15.5% over 20d, so a miss would hit hard.
2. Iran's 7-day Hormuz offer [50], due around Oct 2. If Iran follows through, oil keeps falling and my Energy downgrade holds. If it stalls while the Houthis keep attacking [50][71], crude could quickly revisit $106+.
3. The Fed's path. The October hike my 9/23 read said was priced, plus Warsh's comments [56]. The FOMC date isn't in today's materials. A pause in yields without a dovish shift keeps the avoid on rate-sensitive sectors in place.