Full Scan Market Brief — May 23, 2026

Issue #18 · Clean Sweep Edition · 12 min read

---

1. Three Things

1. QQQ RSI 71.5, XLK RSI 73.0, SMH +39.6% 3M — the AI compute layer has gone full melt-up, but the driver is semis and networking, NOT Mag7: XLC (comm services) is -1.64% 30D while XLK is +15.75% 30D.
2. 12 profit-target exits today — SNOW +41.6%, DDOG +71.7%, ENPH +83.4%, QCOM +85.7%; the paper system harvested a massive gains cycle while the cybersec cohort (FTNT RSI 88.5, CRWD RSI 87.4, PANW RSI 84) races into earnings prints next week — pre-print trim is now mandatory.
3. ARM +48.7% in a single week — chip IP licensing is being re-rated as the royalty stream on every AI accelerator ever deployed; 202 parabolic monsters confirmed; zero capitulation in the bargain bin; and the NVDA Q1 FY27 print on 5/28 is the fulcrum everything hangs on.

---

2. The Big Picture

This is the first clean full-scope scan in weeks, and what it confirms is both reassuring and uncomfortable. Every major index has a golden cross, RSI 60–71, and is within 1.5% of ATH. Bonds are in a death-cross downtrend, gold has lost 14% in three months, and volatility products are -53% from their 52-week high. By every historical measure, this is a maximum-confidence melt-up tape.

Act 1 — The Narrow Leaders. The index-level story looks broad, but one level down it's not. XLK is up 30% in three months while XLC is down. QQQ is up 19.5% 3M while the equal-weight RSP is only up 2.78%. The rally is real — IWM near ATH, small-caps participating — but the engine is a very specific stack: ARM (chip IP), MU/STX/WDC (memory), CRDO/ALAB/LITE (optical networking), and the cybersec cohort (FTNT/CRWD/PANW). Everything else is either pausing or broken. Enterprise SaaS (CRM -35% from ATH, ADBE -42%, WDAY -50%, NOW -52%) is structurally broken with death crosses across the board. The Mag7 hyperscaler complex (MSFT -25% from ATH, META -23%) is nowhere near leading this tape.

Act 2 — The Pre-Print Wall. The entire cybersec cohort is running into a three-print earnings wall: ZS 5/26, PANW 6/2, CRWD 6/3. FTNT already printed and is at RSI 88.5. CRWD at RSI 87.4, -1.7% from ATH, going into its print on 6/3 is textbook sell-the-news risk — not a thesis exit, but a trim event. Insider selling into these names is running $2B+ aggregate. Meanwhile, NVDA at RSI 53.6, cooling to near its 20-day SMA, prints on 5/28 and is the single most important event this cycle: a miss or soft guide is an AI-wide flush; a beat and strong guide is the green light for another leg up across the entire infrastructure stack.

Act 3 — The Opportunity Layer. While the headline tape chases overbought, three quality names hit deep oversold from earnings shocks: WMT RSI 34.1 after a -9.18% single-week drop, DE RSI 32.4 after -17.95% in three months on ag cycle contraction, and INTU RSI 33.6 after a -21.13% single-week earnings collapse. The bargain bin had zero names below RSI 30 before these shocks — these are rare gift setups in a market where everything is extended. All three were bought across multiple strategies today.

---

3. Focus List

NVDA is the most important focus ticker this week. RSI 53.6, +0.3% above SMA20, -8.97% from ATH — this is the "healthy consolidation before the print" pattern. Hold all AI ecosystem positions through 5/28. If NVDA guides conservatively on data-center capex commentary, every AI name gets repriced lower simultaneously.

ARM at +48.7% in seven trading days is the wildest single-week move in the focus list. The IP licensing model means every NVIDIA Blackwell chip, every Qualcomm mobile SoC, every Apple A-series chip pays ARM a royalty — the market is suddenly pricing that as a compounding annuity rather than a semiconductor cyclical. At RSI 76.6 and +108% above SMA200, don't add; hold if you have it.

GOOGL gave back -5.7% this week alone while maintaining its golden cross. The -6.27% from ATH gap and RSI 57.5 make it the cleanest risk/reward in the mega-cap cohort right now — not chasing, not broken.

NET (Cloudflare) is the most important cybersec distinction: RSI 54 while FTNT/CRWD/PANW/ZS run RSI 73–88. Golden cross, +272% three-year return, still has room. When the cybersec earnings season delivers sell-the-news on PANW/CRWD, NET is where you park the capital.

---

4. Sector Scorecard

---

5. Market Vibe

The tape is doing something specific: it's repricing the AI capex cycle as a multi-year compounding annuity instead of a cyclical burst. The numbers that tell this story — MRVL +152% 3M, ARM +148% 3M, INTC +175% 3M, CRDO +76% 3M — are not rotation plays. These are structural re-ratings of companies that sit at the base of the compute stack. When every hyperscaler is adding hundreds of thousands of GPUs per quarter, the IP royalties, the custom silicon design wins, the switch ports, and the optical transceivers all compound. The memory supercycle (MU +79% 3M, STX +100% 3M) is the same story told in DRAM — HBM demand for AI training is not a cyclical bump, it's a structural shift in what memory is for.

What's weird is the divergence within the AI stack. NVDA (RSI 53.6) is cooling while ARM (RSI 76.6) and the cybersec cohort (RSI 84–88) are parabolic. The market is not bidding equally on every AI name — it's making fine distinctions. NVDA is digesting a regulatory-overhang narrative (H20 chip export restrictions to China) while ARM's IP model is being repriced without the same headwinds. The enterprise SaaS layer (ADBE/CRM/WDAY) is being punished precisely because AI threatens to disrupt software licensing models — the market sees that threat and is selling it before the earnings prove it.

The most important thing that doesn't fit the narrative: gold is down 14% in three months while geopolitical risk hasn't actually gone away (contested Iran-deal attribution, ongoing Ukraine, China Taiwan tension). The safe-haven bid has fully evaporated. Either the market is correctly pricing geopolitical tail risk as manageable and contained, or it's not pricing it at all. Given VXX at -53% from its 52-week high, the market is firmly in option B. That asymmetry is building — not a sell signal, but a position-sizing discipline signal for anyone adding new positions.

---

6. Scan Dashboard

6a. The Dashboard

6b. Expanded Dispatches

Cybersecurity — Sell-the-News Setup Forming. The three biggest names in the cohort — ZS (RSI 73.6), PANW (RSI 84), CRWD (RSI 87.4) — are all printing in the next 11 days. FTNT already printed and is at RSI 88.5; the template is "beat and gap up, then stall at extreme RSI." PANW at -0.3% from its 52-week high with RSI 84 and $1.7B in insider selling going into a 6/2 print is the highest sell-the-news risk in this market. The cybersec earnings-acceleration thesis is confirmed — this is a pre-print trim event, not a thesis exit. NET (RSI 54, golden cross) is where to be after the earnings dust settles.

Optical Supply Chain — Parabolic Everywhere. AXTI +395% 3M, SIVEF +373% 3M, ALAB +43% in a single week (RSI 79), LPTH +56% in a single week. The AI datacenter transceiver and photonic interconnect demand story is no longer a research thesis — it's priced into everything simultaneously. The heterogeneous "wave rotation" from LITE/COHR (Wave 1, cooling to rational RSIs 53–60) to ALAB/SMTC (Wave 3, overbought RSI 77–79) is playing out exactly as the perspective predicted. Wave 4 (LWLG, pre-revenue polymer modulators, RSI 48.9) is digesting and remains the speculative binary. ONTO and FORM (RSI 46–49, -10–13% 30D) are the clean pullback entries on the confirmed thesis.

Bargain Bin — A Gift Hiding in Plain Sight. Zero stocks hit RSI 30 in the entire bargain bin universe before this week's earnings shocks — the melt-up had cleared every oversold setup. Then WMT dropped -9.18% in one week (RSI 34.1), INTU collapsed -21.13% in one week (RSI 33.6), and DE extended its decline to RSI 32.4. These are A+ or A quality businesses at genuine discount. WMT bought across three strategies today; DE bought across two; INTU bought by claude-momentum. The caveat: INTU's drop is explicitly AI-disruption-to-TurboTax narrative, not a macro pullback — the thesis has a real structural question embedded. Accumulate WMT and DE with conviction; size INTU smaller.

AI Infrastructure — Two-Speed System. The networking layer (FFIV RSI 81.7, OSS RSI 73.6, EXTR RSI 72.1, ERIC RSI 72.4, NOK RSI 71.5) is running hot — the R27-E adds (EXTR/ERIC/FFIV/ATEN) were exactly right and are now near ATH. The physical infrastructure layer beneath it — aggregates (MLM RSI 27, VMC RSI 31.6), utilities (NEE RSI 38.8), and cooling (ETN, TT pulling back) — is in a hard correction. The paradox: the companies building the backbone for AI datacenters (aggregates, power systems) are the only names not participating in the AI rally. MLM at RSI 27 approaching capitulation is the clearest contrarian setup in this watchlist.

Monster Scan — 202 and Climbing. The count has gone from 191 (R27-A, April 26) to 202 today — steady accumulation of parabolic names, not a spike. New entrants NVTS (+190.9% vs SMA200, GaN/SiC AI power), VPG (+164.5%, force sensing now a datacenter electrical story), and NOK (+113.9%, AI RAN and optical networking) are the most interesting additions. ARM entered the parabolic tier at +108.4% vs SMA200. The overbought cluster (42 names RSI > 70) is at the high end of historical range — not a sell signal, but the position-sizing table is telling you to be smaller on new entries.

---

7. The Front Page — Deep Research

7a. Perspective Dispatches

AI Power Bottleneck [id: ai-power-bottleneck] — The watts-per-rack thesis continues to play through the hardware layer. BE +88.7% 3M (RSI 62.8) is the cleanest hold in the cohort — fuel cell datacenters as baseload backup, not overbought. CLSK and RIOT are running the miner-to-AI-compute pivot: both up 56–63% 3M. The most important data point this cycle is NVDA Q1 FY27 on 5/28 — look for power infrastructure commentary in the earnings call. GEV/gas turbine backlog language matters more than the EPS beat.

AI Power Delivery [id: ai-power-delivery] — The 30% parasitic loss thesis is finding its expression in WOLF (+153.7% 30D, RSI 76.3 — SiC GaN parabola), NVTS (+52.5% 7D, RSI 72.9 — GaN/SiC AI power), and the challenger cohort (AOSL/VICR/NVTS/POWI). The perspective was refreshed 5/22 as "stale-right, not stale-wrong" — MPWR Q1 resolved positively, ETN remains the cohort laggard supporting the incumbent-bear leg. WOLF's equity re-rate is balance-sheet repair rather than substrate-share proof; the asymmetry tightens onto challengers vs incumbents.

Optical Supercycle [id: optical-supercycle] — Wave 1 (LITE/COHR/CIEN) passed the demand-confirmation test: COHR Q3 FY2026 record revenue $1.81B (+21% YoY), Datacenter segment $1.4B (+37% YoY). Wave 3 (TSEM RSI 67.8, ALAB RSI 79) is the current momentum layer — ALAB +43% this week alone. Wave 4 (LWLG RSI 48.9) is digesting its 3M +264% move below SMA20, still above SMA50. The ONTO/FORM pullbacks (RSI 46–49, -10–13% 30D) are the cleanest wave-1.5 accumulation points.

Risk-On Tech Rotation [id: war-ends-playbook] — The cybersec leg (added R15) is the active trade now: FTNT RSI 88.5, CRWD RSI 87.4, PANW RSI 84. The watch-for trigger for this leg is "catalyst-exhaustion" — all three June earnings beat but cohort fails new highs within 5 trading days post-prints. DZ Bank's FTNT downgrade (5/19) is the leading edge of that signal. NVDA Q1 FY27 commentary on 5/28 is the other watch event.

Gold Anomaly [id: gold-crash] — GLD $413.82 (RSI 39.6), now -18.81% from ATH, approaching the SMA200 at $400.12. The structural watch-for trigger has arrived: GLD < $410 = full breakdown per the perspective log. The safe-haven failure thesis continues to confirm — four consecutive geopolitical triggers (Iran narrative, late-May countdown, oil pullback) and gold didn't bid. Miners (GDX, GDXJ) leading spot lower at -22–24% 3M is the worst possible signal for a recovery.

Memory Supercycle [id: memory-supercycle] — Born 5/19 from the verified memory supply-shock research note. MU +55.9% 30D, STX +38.3% 30D, WDC +20.1% 30D — all golden-crossed, all near ATH. The Samsung NSEU strike (started 5/21, ~45k workers) is a transient catalyst layered on top of the structural DRAM/NAND/HBM tightness. TrendForce Q2 2026 contract prices DRAM +58–63%, NAND +70–75% QoQ. Memory is the highest-quality fundamental buy in AI right now — unlike cyber, the fundamentals are clearly inflecting.

US Energy Dominance [id: us-energy-dominance] — USO -4.98% 7D on the contested Iran-deal narrative (TheShortBear 5/22: "no deal close"). Services (SLB RSI 62.2, HAL, OIH) outperforming upstream producers on every timeframe — SLB is essentially at ATH while XOM/CVX are -11–12% off theirs. The Canadian premium (ATH.TO +156% 1Y) and Buffett's OXY accumulation continue as the most durable expressions of this thesis.

Gulf Infrastructure Strike [id: gulf-infrastructure-strike] — Ras Laffan (Qatar) shut 2026-03-02 cutting ~30% of global helium supply. The August/September 2026 stockpile depletion deadline is the binding event. APD Doe Canyon moat (+LIN) is the working US-side leg; Asian semiconductor exposure (TSM's 64% Qatar helium dependency) remains underpriced by the tape.

Oil $200 Scenario [id: oil-200-scenario] — Demoted to tail-risk. USO > $130 trigger fired (extended to 7/1 auto-retire), but the thesis remains medium-priority only. The contested Iran-deal narrative is the live watch: if no deal materializes, the geopolitical risk premium removed from USO this week was premature.

Crypto-Geopolitics [id: crypto-geopolitics] — BTC at $74,432 (RSI 37.3) is below the April trigger price ($75K) and without RSI confirmation. But the miner decoupling is the real thesis: RIOT/MARA/CLSK +56–75% 3M vs BTC +15% 3M confirms the "miners converting stranded power into AI compute revenue" story. BTC < $70K would be the thesis invalidation. ETH at RSI 28.7 approaching capitulation; ETH RSI 25 = staged entry signal.

Food Security Cascade [id: food-security-cascade] — Stage 2 (fertilizer, CF -11% 30D) consolidating as nat gas weakness provides input cost relief — the physics says this is right. Stage 3 (commodity leg, WEAT +12% 3M, SOYB near ATH) is doing the heavy lifting. Stage 4 (MOS RSI death cross) still locked at RSI 37. DE at RSI 32.4 is now a direct food-security candidate as ag equipment cycle turns.

7b. Monster Watch

202 parabolic monsters confirmed, up from 191 five weeks ago. The dominant theme is AI networking stack — ARM, ALAB, CRDO, NOK, MRVL, FFIV, EXTR, ERIC all in monster tier now. Memory cluster (MU/STX/WDC/SNDK) is the strongest fundamental sub-group. The 42 overbought names (RSI > 70) represent the trim-or-hold zone; best healthy entries on pullbacks are CRDO, CIEN, NBIS, BE, and WDC.

7c. Deep Dive Spotlight

No new deep dives were written this scan (deep dives are run manually, not during full scans). For deep-dive coverage on any ticker that surfaced in the scans above — ARM, NVTS, NOK, CRDO, MLM, WMT, INTU — use /deep-dive TICKER.

7d. Sector Rotation Radar — Candidate Themes

---

8. Key Signals

Gold / Commodities. GLD $413.82, RSI 39.6, -18.81% from ATH, approaching SMA200 at $400.12. This is the moment the gold-crash perspective named: four consecutive geopolitical triggers and gold didn't bid any of them. GDX/GDXJ miners down -22–24% 3M while gold spot is only -14% — miners leading the metal lower is the worst structural signal. If GLD closes below $400, the long-term trend is broken and the "accumulate on weakness" thesis gets suspended. GDRZF royalty divergence (+16.25% 30D vs GDX -7.78%) shows cash flows are healthier at the royalty layer than the mining layer.

Insider Buying. The single clearest signal is TSLA: $206.3M net insider buying, RSI 58.4, +14% 30D — the largest absolute insider buy in a 60-stock universe. CRSP ($28.1M net buy into -8.74% 30D) and ASAN ($8.4M) are the only other net buyers. The dominant story is the other direction: cybersec insiders distributing $2B+ aggregate (PANW -$1.7B, CRWD -$433M, DDOG -$586M) exactly at RSI 84–87. The dual-risk flag (RSI > 80 + net insider selling) on PANW, CRWD, and DDOG is the most credible distribution signal in the market right now.

Crypto. BTC $74,432 (RSI 37.3, -41% from ATH) is NOT confirming the equities melt-up — the two markets are moving independently. ETH RSI 28.7 is approaching true capitulation (RSI 25 = staged entry signal). The story worth tracking is the miner decoupling: CLSK +63% 3M, MARA +75% 3M, RIOT +56% 3M vs BTC +15% 3M. Miners are being re-rated on hash efficiency and AI compute optionality, not BTC price appreciation. SEC approved Nasdaq Bitcoin index options (medium-term constructive, H2 2026 rollout). Bought MARA and BITQ in yolo/bench-signals today to capture the miner re-rating.

Russell Rebalance. GOOGL and AMD move from value to growth indices on June 26, 2026. This is a positioning catalyst independent of fundamentals — passive funds tracking growth indices must add; value funds must reduce. Watch for positioning into late June.

---

9. The Wild & Whacky

• ARM +48.7% in a single week. One week. A $300 stock. This is what "re-rating" looks like in real time — the market decided chip IP licensing is a compounding annuity, not a cyclical, and repriced accordingly. Still RSI 76.6 at +108% above SMA200. Do not chase.
• INTU -21.13% in a single week. The steepest single-week decline in the bargain-bin watchlist history. The AI-disruption-to-TurboTax narrative is the market's verdict — not just a guidance miss. Now -60.68% from its 52-week high. Bought at the lows today.
• VPG (a force-sensing company) is +96.9% in 30 days. VPG makes precision resistors and strain gauges. It's now classified as a "datacenter electrical components winner" and is a monster at +164.5% above SMA200. Sometimes the market finds signal in the strangest places.
• Inverse ETFs at RSI 26–28. SOXS RSI 28, SQQQ RSI 28.1, TECS RSI 26.1 — bear capitulation is essentially complete. The cost of hedging against semi/tech/QQQ declines is at a multi-year low. This is not a reason to buy hedges, but if you were going to own some, now is when the math works.
• BB (BlackBerry) RSI 85.1, +138% 3M. BlackBerry. The phone company. +21.7% in a single week, RSI 85. It's technically in the cybersec watchlist (IoT security division), but the move has the unmistakable signature of a momentum squeeze on a recognizable brand name. Not a fundamental play.
• AXTI +395% in three months. AXT Inc makes GaAs and InP compound semiconductor substrates for photonic devices. The stock has gone from $28 to $141 in 90 days on AI optical transceiver demand. Still not RSI-overbought at 68.1 — the photonic substrate story is that early.
• Gold and equities both ignoring each other. GLD -14% 3M while SPY +9.6% 3M. Equities at ATH with no inflation bid in gold = "no fear, no defense" maximally extended. This co-movement (or rather, anti-co-movement) historically resolves in one direction: either equities catch down or gold stabilizes and moves up. The SMA200 at $400 is the tell.
• NVDA RSI 53.6 while ARM RSI 76.6. NVDA makes the actual chips; ARM licenses the IP that goes into those chips. ARM is more extended than NVDA right now. The royalty layer is being valued above the manufacturer layer. That is unusual.

---

10. Paper Trade Report Card

Danger zones: yolo is carrying NUGT (-16.2%) and GDX (-9.9%) — gold miner positions from the first half of 2026 that haven't recovered. The gold-crash perspective is the backdrop; if GLD breaks $400, these need reassessment. ABT in bench-signals is at -13.8%, approaching its -15% stop.

Best thesis working: SNOW across all four strategies — bought at RSI 21 in April, profit-targeted at +41.6% today. This was a cross-strategy conviction trade that paid in full.

Real-money advisories (user decision required):

• MDT below its stop level ($78.60 vs $83 stop) — thesis freshness check needed
• BTC/GBTC position — RSI 37, -41% from ATH; watch for RSI 25–30 as staged-entry signal

---

11. What I'd Tell a Friend

The hardest thing to do right now is resist the temptation to chase what's already run. FTNT at RSI 88.5, ARM at RSI 76.6, ALAB at RSI 79 — these are great companies in a great cycle, but "great company" and "great entry" are not the same thing at RSI 80+. The better plays are the names in the same AI infrastructure cycle that haven't run yet, or the quality companies that got sold on earnings shocks.

My highest-conviction entries right now:

1. NVDA — accumulate at current levels ($215) before the 5/28 print. RSI 53.6, +0.3% above SMA20, -9% from ATH. This is the "healthy consolidation before the catalyst" pattern. If you don't own NVDA going into Q1 FY27 earnings, the risk is asymmetric to the upside. Size it for a potential -10% gap if the guide disappoints. Entry zone: $210–$220.

2. WMT — accumulate in tranches ($118–$122). RSI 34.1, -9.18% single-week shock. Walmart is one of the two or three most defensible businesses in consumer retail — pricing power, grocery, private label, Walmart+. This is not a structural break, it's an earnings-surprise discount. Target: $156 (+30%). Stop: $102 (-15%).

3. NET — add on any pullback to $205–$210. RSI 54.5, the only large-cap cybersec name that is NOT overbought. When PANW and CRWD print sell-the-news reactions in June, NET is where that capital will rotate. -16.86% from ATH with golden cross and RSI 54 = structural runway.

4. GOOGL — light accumulate at $380. RSI 57.5, -6.27% from ATH, golden cross, -5.7% this week. This is the biggest Mag7 discount with intact structure. The advertising revenue cycle, Gemini optionality, and YouTube monetization are not broken. The -5.7% this week is noise; the golden cross is signal.

5. BITQ — crypto miner basket ($25–$28). RSI 58.1, golden cross. If you want crypto exposure, don't buy BTC at RSI 37 — buy the miners re-rating on AI compute optionality. BITQ gives diversified exposure across RIOT, MARA, CLSK, HUT without single-name risk.

Watch for: NVDA's 5/28 earnings call. The question is not whether they beat revenue — it's whether Huang's commentary on AI capex runway (2026, 2027, 2028 visibility), data center power commentary, and H20/export-control headwinds either reaffirms or resets the AI infrastructure multi-year thesis. Every single name in sections 6, 7, and 8 above trades on the answer.

---

12. Active Perspectives

---

13. Scan Summary

R30-A — First clean full Phase 0–9 scan of the cycle. Data: 5/22–5/23/2026.

Scans completed (29 total): market-pulse · etf-universe · ai-scan · ai-infrastructure · optical-supply-chain · cybersec · bargain-bin · insider-scan · tech-insider-buys · macro-commodities · defense-contractors · geopolitical-risk · crypto-scan · ev-clean-energy · airlines · monster-scan · biotech-scan · healthcare-scan · consumer-scan · retail-scan · cloud-etfs · defensive-scan · nvda-ecosystem · mag7 · biotech · space · web-cloud · energy · financials

Candidate themes refreshed: 11 (ai-infra, biotech, cloud-saas, consumer, crypto, energy, mag7, memory, retail, space, web-cloud)

Deep dives written this scan: 0 (manual — use /deep-dive TICKER)

Active perspectives: 11 (ai-power-bottleneck, ai-power-delivery, optical-supercycle, war-ends-playbook, gold-crash, memory-supercycle, crypto-geopolitics, us-energy-dominance, gulf-infrastructure-strike, food-security-cascade, oil-200-scenario)

Perspectives closed: protein-economy (closed 5/19 — thesis failed, BRBR stop triggered, GLP-1 compounding channel shut)

Paper trades: 12 exits (SNOW +41.6%, DDOG +71.7%, ENPH +83.4%, QCOM +85.7%, BRBR -40.6% stop) · 12 buys (WMT/DE/INTU/ANET/MARA/BITQ/LULU/DRAM across 4 strategies)

Critical market freshness: PASSED — SPY/QQQ/IWM/RSP/XLK/SMH/VTI/DIA/GLD/USO/TLT all fresh 2026-05-22.

Russell rebalance: GOOGL/AMD value→growth on 2026-06-26.

Previous brief: May 1, 2026

Sources

Validated daily summaries and OHLC, close of 2026-05-23.

• Fundamentals figures: company-reported results (quarterly/annual filings) as available at the artifact date; predates the desk's EDGAR reconciliation gate — figures not re-verified after publication.