IGV Breaks Toward 117; CRM Still Capped at 277
US Stocks · Options · News · Views
The Averages Were Green; Ten Sectors Bled
Thursday, August 27, the four averages finished green. Nasdaq led. The Dow gained 0.2% to 53,569. SPX rose 0.7% to 7,731. Nasdaq rose 1.6% to 26,541. NVDA closed 227.98, up 8.74%. SOX closed 11,882, up 2.33%, with the 25-name basket following. Software was dragged up by CRM, up 22.58%, and CRWD, up about 20%.
The averages were green. The vote was not. About 152 SPX names rose, 349 fell. Of eleven sectors, information technology was the only one up, about 3%. The other ten all fell. Semiconductors and software had been on opposite ends of a seesaw for nearly a year. Thursday they sat on the same side. The rest of the tape sat on the other. Flow was clear: keep bidding AI; software that had been priced for an AI kill-shot got paid if the turn is already lifting revenue. CRM and CRWD put AI strategy and commercialization on the table. The tape accepted it. AI and software can be rivals or partners. Who keeps up is a management-speed problem.
Macro did not open a second story. Initial claims fell from 206,000 to 203,000, under about 208,000. The four-week average ticked up to 205,500. Continuing claims fell from 1.799 million to 1.778 million, under about 1.79 million. Rate books still wait on Friday’s Warsh speech. The real pricing was the software break, the chip squeeze, and the MRVL print after the close.
Marvell Cleared the Print; After Hours Gave the Multiple Back
MRVL closed 241.45 cash, after trading from about 253 down to 240. After the print it sat near 222, down almost 8%. Second-quarter adjusted EPS was $0.94 versus $0.67 a year ago, a tick over about $0.93. Revenue was $2.739 billion, up 37% year over year, a tick over about $2.72 billion. The numbers cleared. The guide ticked up: next-quarter adjusted EPS $1.05 to $1.15, revenue midpoint $3.15 billion, versus about $1.08 and $3.04 billion. Further out, management lifted fiscal 2028 revenue to about $18 billion from about $16.5 billion. The Google piece does not become a material slice of total revenue until fiscal 2029 — almost two years.
A small beat and a small raise do not turn rich into cheap. After-hours 224 is closer to a 216 midpoint. Long-term buys sit near 211, the 188 to 211 shelf; a wider margin of safety is 160 to 172. Both overlap technically and on the fundamentals. Overhead 253 to 272 and 278 to 318 are still heavy. Trims already taken at the high stay taken. Cutting here does not add much, because 211 is a buy zone again. Google still waits. Mix and margin sit inside the custom-silicon shift. After hours paid the premium back first.
CrowdStrike Is an ARR Story; Salesforce Is Still Capped at 277
CRM closed 252.05, up 22.58%, high 254.48. 277 did not print. The 52-week high at 269 is still above. Volume was abnormal. With the rest of the tape drained and only semis and software ripping, part of the bid looks like a short cover. Software positioning was already light; shorts had come down and were still high. There is no positioning tape to prove it. Treat it as a guess. After the cover is done, whether 277 clears decides if spring actually arrived. If it fails, look for stall and topping over the next few sessions.
CRWD jumped about 20% to 227.96. Second-quarter adjusted EPS was 31 cents on $1.47 billion of revenue, up 26% year over year, a fifth straight quarter of acceleration, above 29 cents and $1.44 billion. Net new ARR was $332.8 million, up 51% year over year, well above the high end of guide. Total ARR was $5.84 billion, up 25%, a fourth straight quarter of acceleration. Once the subscription is signed, it stacks. The guide ticked up: fiscal 2027 revenue $5.991 billion to $6.011 billion, adjusted EPS $1.25 to $1.26, versus $5.915 billion to $5.959 billion and $1.22 to $1.24. EPS did not add much. The ARR stack is what wrote the path. The more enterprise AI gets deployed, the more holes appear, and the more cybersecurity work shows up. These names also use AI against AI.
Near-term valuation is stretched. A long-term book does not buy this. The tape is a clean right side: no overhead, earnings are good, and without a real negative it is hard to go deep. Do not expect 20% days on repeat. When the cover runs out of ammo, the push fades. Swing trades wait for a topping pattern. Long-term size waits for the next raise and the next ARR step-up. PLTR, NVDA, and AMD have all lived the stretch where raises keep coming, the market’s appetite grows, and a raise that is not large enough starts to pay the multiple back.
IGV Broke First; Software Does Not Lift as a Bloc
IGV closed 110.32, up 7.74%. It opened about 105.69, cleared 107 and 108, could not be sold on the dip, and stacked higher on volume. Cover plus right-side chase. The shape is clean. Once the range high breaks, the map is a new high. The 52-week high is about 118. The rule is hard: any dip that holds 101 does not break the new-high structure. An open or close through 101 puts it back in the 90 to 117 range. About seven points to the high. Semiconductors are still grinding downstairs. Software tested the ceiling three times. This time it went through first.
Software components did not fly together. Location matters. Some names were up more than 10%. Some barely moved. Passive money will not lift the hardest chart just because the group is green. ORCL has too much overhead and support too close — a tight box that cannot lead. PLTR closed 185.93, high 186.86. Through 192 the structure is a new high, with only small resistance left; the stop is 175. If money is going to work software, that path is simpler than Oracle’s. INTU came off 346 to 363, with 321 as support, stuck in the door. APP broke down, stacked resistance overhead, and needs a long coil. ADBE is climbing into 322 to 361, the real test. PANW closed 382.85, up about 13%, same as CRWD — pinned to highs, no overhead, forward-value cap around 163 against a doubled price. FTNT closed 172.78, also a right-side high; the chase window is gone. NOW closed 138.43, up about 10%. After the 120 break it held 110 and is still right-side. First map is 147 to 159; through that, 180 to 193. A later high-30s-to-40s mean reversion waits on a fiscal 2027-end raise of about 15% to 20%, toward 220 to 240. Clear 150 and 159 first. SNPS closed 464.89. Overhead 488 to 527 is heavy and close.
The swing list that still matters: NOW, the three cybersecurity names, whether CRM takes 277, and PLTR a few points from 192.
Nvidia Missed by Two Cents; 514 Has Not Traded
NVDA printed huge volume. Cover was part of it. Fiscal 2028 revenue growth around 70% sat the public shorts down. A new high is not far. Do not expect stacked 8%, 5%, 7% days. At this cap it climbs stairs: a small dip, a pause, then another step. Close 227.98. 228 missed by two cents. Two cents is still not a break. Overhead is light. Whether it rejects, retests, and becomes medium resistance is later. Daily, weekly, and monthly still point up. Holders keep it. There is no heavier cap above. Get through 228 first.
MSFT closed 505.06, high 506. The break line was tightened from 516 to 514. Through it, 600 is the map. The forward book: June 2027 458 to 614, midpoint 536; June 2028 540 to 724, midpoint 632. Into 2027, on the current AI-monetization path, 600 is not a stretch. Pinning a date to a price is not a trade. Hold the stock. Do not map this onto option expiry.
Friday is Warsh. Further out, September and October are historically noisy, including around midterms. Books that slept through August should wake up. Near term, three items: whether IGV holds 101, whether CRM takes 277, and whether NVDA still owes those two cents. MRVL already printed the next bar’s start: 211 is where a long-term bid belongs again.