Five tech names, three of them firing this week, and the referee says sell the spike, not the story. MRVL and CRM print tonight (5/27 AMC), DELL tomorrow (5/28) — all three carry Fade-Risk ≥ 6.8 against a market where dealer gamma is +0.82M positive and the regime is still STAGFLATION_HOT, the exact recipe that round-tripped DE, MDB and ZS (now 3/3). The single new input that matters most: HSBC upgraded MRVL to Buy with a Street-high $300 target into the print — bullish, but the kind of pre-print euphoria that tends to be the top of the pop, not the start of a leg. CRM is the dangerous one: an 88% beat rate that doesn't matter, because BofA reinstated Underperform on 5/26 calling AI a "major risk" and Agentforce a non-starter — a textbook beat-and-fade setup. AVGO (6/3) remains the one genuine hold-through carry; honor its untouched $411 stop and do not repeat the day-1 tightening that cost A$531 last week.
The trigger prompt's position line reads "CHECK THE JOURNAL" — no hard-coded tickers, so no conflict to resolve. The PDS_Trade_Journal.md OPEN POSITIONS header (mark 2026-05-26 close) is the live book and was marked verbatim: MRVL 41 · ORCL 83 · DELL 31 · CRM 48 · AVGO 32. ZS (exited 5/26 @179.07) and MDB (exited 5/26 @313.80) are closed and not carried.
One mark divergence flagged: the journal header marks AVGO at 421.88 (sourced "UW 1D close series"); the Phase-3-mandated get_ticker_ohlc_latest_or_date regular-session close is 422.01. This run marks 422.01 (the OHLC endpoint), a A$5.8 improvement on AVGO's unrealized loss. Reply to re-mark if you want the 421.88 basis retained.
Every price, IV, flow, macro print and analyst action below was fetched in THIS session (UW OHLC/greeks/earnings/ratings, Benzinga, web). No data point is reused from prior reports or memory — the prior report is the conviction-evolution anchor only. Data gaps disclosed in the footer.
| Instrument | Level | Δ | Read |
|---|---|---|---|
| S&P 500 | 7,519.12 | +0.61% | record close, chip-led |
| Nasdaq Comp | 26,656.18 | +1.19% | record, AI surge |
| Dow | 50,461.68 | −0.23% | healthcare/energy drag |
| Russell 2000 | ~2,890 | +1.77% | >2,900 record intraday |
| VIX | ~16.6 | −0.7% | calm; futures 18.7 |
| UST 10Y | ~4.48–4.51% | −9bp | cooled |
| UST 30Y | ~5.03% | −5bp | |
| WTI | ~$92–93 | vol | Mid-East war premium |
| Brent | ~$96 | — | |
| Gold | ~$4,536–4,570 | −0.5% | near highs |
| DXY | ~99.0 | + | UUP proxy (DXY premium-locked) |
| BTC | ~$76,700 | −0.4% | −13.7% YTD, AI-capital rotation |
Futures premkt 5/27 (UW 00:16 UTC): SPX +0.09% · NDX +0.09% · RUT +0.24% · DJI +0.12% · VIX 18.70 — mildly green, calm. Micron +19.3% topped $1T on UBS upgrade = the day's chip catalyst.
Drivers (cited): (a) inflation expectations surging — UMich sentiment at a record-low 44.8, CB Consumer Confidence 93.1 (beat 92 est but down from 93.8, present-situation −3.2 to 121.2) with the Conference Board explicitly blaming "the inflationary impacts of the war in the Middle East" (5/26); (b) growth/confidence softening while equities melt up on AI; (c) energy still war-bid (WTI ~$92, gas at highest Memorial Day since 2022). Yields actually cooled 9bp — the one non-stagflationary note — but the inflation-expectation spike dominates the bucket.
Fed: Kevin Warsh began as Fed Chair the week of 5/26; FedWatch prices no June cut. The week's heavy data (PCE, claims, durable goods, GDP 2nd) all land 5/28 BMO — AFTER the MRVL/CRM prints and the morning of DELL's. The sequencing matters: tonight's two prints face a market that has not yet seen the inflation data, so a hot 5/28 PCE could retroactively punish any name that gapped up on 5/27. That is itself a reason to pop-sell into strength rather than carry the binary gap through both the print and the macro release.
Why the bucket holds despite cooling yields: the 9bp drop in the 10Y would, in isolation, argue for a softer reading. But the classifier weights the inflation-expectation channel — UMich 44.8 record low, year-ahead expectations ~4.8%, gas at a multi-year Memorial-Day high — above a single duration move that more likely reflects growth fear (stagflationary in itself) than genuine disinflation. STAGFLATION_HOT stands, de-intensifying only at the surface.
SPX dealer gamma +0.82M net positive (UW greeks 5/26: call γ 3.74M, put γ −2.91M) — vol-suppressive, mean-reverting, rallies fade-prone. Eased slightly from +1.11M on 5/22 but still firmly positive. Combined with STAGFLATION_HOT and a 3/3 peer-fade record this cycle, this biases every name's Fade-Risk up and is the spine of tonight's pop-sell calls.
Filtered OUT: Ferrari Luce EV launch (RACE), Apple iPhone 17e Japan sales, Vatican AI ethics, newsroom data leak, CA governor race, Bitcoin-vs-AI think-pieces — no thesis impact on the held book.
| Tkr | Sh | Entry | 5/26 Close | P&L USD | P&L AUD | % to Stop | Stop · TP src | Next ER |
|---|---|---|---|---|---|---|---|---|
| MRVL | 41 | 168.53 | 208.26 | +1,628.93 | +2,277.3 | +6.80% | 195.00 (5/26 raise) | 5/27 AMC |
| ORCL | 83 | 185.24 | 193.06 | +649.06 | +907.4 | +2.16% ⚠ | 188.98 | ~6/10 |
| DELL | 31 | 244.73 | 305.08 | +1,871.00 | +2,615.7 | +5.20% | 290.00 (5/26 raise) | 5/28 AMC |
| CRM | 48 | 178.07 | 179.08 | +48.48 | +67.8 | +4.76% | 170.95 | 5/27 AMC |
| AVGO | 32 | 427.50 | 422.01 | −175.68 | −245.6 | +2.68% ⚠ | 411.00 | 6/3 AMC |
| OPEN UNREALIZED | +4,021.80 | +5,622.5 | ||||||
Per-position AUD sum = 2,277.3 + 907.4 + 2,615.7 + 67.8 − 245.6 = +A$5,622.6 ≈ header +A$5,622.5 (<A$1 rounding). Realized cum +A$3,205.7 (unchanged — no new exits). NET TOTAL +A$8,828.2. Open is +A$5.8 vs the journal header's +A$5,616.7 purely from the AVGO mark divergence (422.01 vs 421.88).
ORCL +2.16% above 188.98 and AVGO +2.68% above 411.00 are the two tightest — next stop-out candidates on any fade. MRVL (+6.80%), CRM (+4.76%), DELL (+5.20%) have room. AVGO discipline rule ACTIVE (banked 5/26): do NOT tighten the $411 stop on day-1 negative drift — the prior 76-share AVGO loss (−A$531, 5/22) came from exactly that. No position within 2% of a TP1.
| Tkr | Hold-through | Pop-sell | Fade-Risk | Primary track |
|---|---|---|---|---|
| MRVL | 6.9 | 8.5 | 7.0 | POP-SELL |
| CRM | 7.4 | 7.6 | 6.8 | POP-SELL beat-and-fade |
| DELL | 6.9 | 7.4 | 7.0 | POP-SELL |
| ORCL | 6.3 | 8.0 | 5.0 | SPLIT (stop-proximity, not ER) |
| AVGO | 7.9 | 8.1 | 4.3 | HOLD-THROUGH OK |
| Signal | MRVL | CRM | DELL | ORCL | AVGO |
|---|---|---|---|---|---|
| 1 · Options skew† | +0.3 | −0.3 | +0.2 | −0.4 | +0.2 |
| 2 · Term structure† | −0.5 | −0.3 | −0.5 | 0 | −0.2 |
| 3 · Whisper / ESP | +0.3 | +0.4 | +0.2 | 0 | +0.4 |
| 4 · PEAD history | 0 | +0.2 | +0.2 | +0.2 | +0.1 |
| 5 · Sector RS 1M vs SPX | +0.5 | −0.6 | +0.5 | +0.1 | +0.4 |
| 6 · Peer read-through† | +0.4 | −0.5 | +0.3 | 0 | +0.4 |
| 7 · Tariff / China exp. | −0.2 | 0 | −0.2 | −0.1 | −0.2 |
| 8 · Base A–G alignment | +0.3 | +0.2 | +0.2 | +0.1 | +0.4 |
| T2 net | +1.4 | −0.9 | +0.9 | 0.0 | +1.5 |
| Fade-risk lean (1·2·6†) | fade | fade | fade | neutral | mild fade |
† Signals 1, 2, 6 feed the Fade-Risk referee (not double-counted into conviction). CRM's negative T2 net (−0.9) is the only outright bearish overlay — AI-risk de-rating + software RS weakness — reinforcing the beat-and-fade call. AVGO's +1.5 is the strongest, consistent with the lone hold-through.
| Tkr | Prior (primary) | Today (primary) | Δ | New input / status |
|---|---|---|---|---|
| MRVL | 8.2 pop | 8.5 pop · 6.9 hold | +0.3 | CITED — HSBC upgrade Buy PT300 (5/26, new 24h) |
| CRM | 7.0 pop | 7.6 pop · 7.4 hold | +0.6 | CITED — BofA reinstate Underperform 160 + Agentforce-no-traction (5/26) raises fade/pop |
| DELL | 8.0 pop | 7.4 pop · 6.9 hold | −0.6 | CITED — ran +5.2% to close 305.08, now above all bull PTs = more priced-in; soft-Q1 reminder |
| ORCL | 5.9 hold | 6.3 hold · 8.0 pop | +0.4 | CAPPED — price +2.2% reg & net_prem −13.3M are real; re-weight capped at 0.3, operative 6.2 |
| AVGO | 8.x pop/run-up | 8.1 pop · 7.9 hold | ~flat | run-up carry intact; faded mildly with tape, within 3% stop |
| Risk | Status | Note |
|---|---|---|
| AI-capex concentration (MRVL+DELL+AVGO, 3 prints 5/27→6/3) | STILL ACTIVE | Rebuilt by the AVGO add; mitigant = MRVL/DELL are pop-sells exiting into strength, de-clustering by 5/28 |
| High-Fade-Risk = sell-the-print pattern (3/3: DE, MDB, ZS) | STILL ACTIVE | Governing rule; all three imminent names carry FR ≥ 6.8 |
| ORCL: fraud class actions / $50B capex / Schiff bill | STILL ACTIVE | No resolving data; + new stop-proximity flag +2.16% |
| ORCL net_premium bearish | STILL ACTIVE | −13.3M on 5/26 (UW) — deepened |
| Correlation data gap (UW get_correlations noisy) | STILL ACTIVE | Sector-reasoned proxy used (see §9) |
| AVGO underwater (−1.3%), within 3% stop | STILL ACTIVE | Discipline: do not tighten the $411 stop on day-1 drift |
| Market fade-risk bias HIGH (positive gamma) | STILL ACTIVE | +0.82M, eased from +1.11M 5/22 — partially de-intensifying, not resolved |
| Infra: DXY quote premium-locked (UUP proxy); FMP treasury 404 (TLT proxy) | STILL ACTIVE | Data-source workarounds in place |
| STAGFLATION_HOT macro (war energy shock, record-low UMich) | STILL ACTIVE | Yields cooled 9bp = mild surface easing; core inflation-expectation spike unresolved |
Clusters: AI-capex = MRVL + DELL + AVGO (high pairwise corr, est >0.7); SaaS = CRM (standalone, AI-risk de-rating actually decorrelates it); ORCL standalone (enterprise/cloud). With the AI-capex trio >0.7, the dominant-cluster rule applies: eff ≈ max + 0.1×(sum−max).
Pop-sell sum (5) = 8.5+7.6+7.4+8.0+8.1 = 39.6 · max 8.5 · AI-capex trio eff = 8.5 + 0.1×(8.5+7.4+8.1−8.5) = 8.5 + 1.55 = 10.05→cap 8.5 for the trio; + CRM/ORCL near-uncorrelated add ≈ 0.5×(7.6+8.0) = 7.8. Adjusted book ≈ 16.3 vs raw 39.6 = 41%. <80% → CONCENTRATION FLAG.
Mitigant unchanged: 3 of 5 (MRVL, DELL — and ORCL on stop) are exiting into strength within 3 sessions, mechanically shrinking and de-clustering the book. AVGO is the deliberate single hold-through carry.
| Scenario | P | Path | Book action |
|---|---|---|---|
| Bull — beats spike, sell into strength | 35% | MRVL/CRM/DELL pop on beats; gamma fades them intraday but first-min spike is sellable | Pop-sell all three into the spike; bank ~A$5–6k realized; AVGO rides |
| Base — split / beat-and-fade | 45% | Beats but in-line guides; spike round-trips (ZS shape), esp CRM; positive gamma caps follow-through | Exit on first spike is the win; hold-through would be the loss. Discipline = sell the pop, don't wait |
| Bear — guide miss / macro shock | 20% | Light guide gaps a name −10/−16% (ZS/MDB precedent); or hot 5/28 PCE breaks the melt-up | Pre-print trims already de-risked; ORCL/AVGO stops absorb; buffer A$8,857 intact |