Intelligence Synthesis — 2026-04-24

Global Narrative

Today's signal ties to a single regime: stagflationary undertow (CPI re-accelerating 2.4%→3.3% MoM, GDP stalling 4.4%→0.5% annualized) layered against a hawkish Fed-chair structural shift (Warsh confirmation 94.4% Polymarket), while equities sit at exact 160d window highs (SPY 710.14, HYG 80.65) and credit complacency is extreme (HY OAS 2.84%). The portfolio is positioned for AI-capex bifurcation — long the silicon/power layer (NVDA/MU/VRT/AMD), short the application-SaaS layer (WDAY/DOCU/TEAM) — with a duration short via TLT and a credit-complacency short via HYG as the macro hedge that monetizes the same higher-for-longer thesis. The single highest near-term risk is the Apr 29 hyperscaler quad (MSFT/GOOG/AMZN/META) which can simultaneously validate the long-hardware leg and crush the short-SaaS leg via covering rally — the integrator explicitly warns these are not independent bets.

Macro Regime

• Regime: neutral with stagflationary risk building (10d duration); risk_on gate OPEN, long-duration gate CLOSED.
• Drivers: CPI 3.29% YoY (re-accelerating), GDP 0.5% annualized (stall), HY OAS 2.84% (complacent vs 26% recession prob), Warsh 94.4%, FOMC May 6 hold 99.45%.
• Portfolio beta implication: trim gross; pair longs with TLT/HYG shorts as carry hedge; IGV +8.6% 5d covering rally is near-term headwind to short-SaaS adds — wait for exhaustion above 88-90.

Top Trade Setups

Top trade setups from the Claude-only baseline synthesis. These are the trades the synthesis-v2 composer emitted before any multi-model arbitration; compare against the multi-model arbiter variant for divergences.

Exit rules — applied to every card below:

1. Price target hit — close the trade. (green Target price on each card)
2. Price stop hit — close the trade. (red Stop price on each card)
3. Time stop (Invalidate date) — close the trade regardless of price. This is the hard deadline beyond which the thesis is no longer valid.
4. Expected-path lag — at each waypoint (5d / 10d / 20d / target) on each card, compare the actual market price to the expected price. If realized return is materially behind the expected path (e.g., <25% of expected at that horizon), reduce size or exit early. Lagging trajectory is an early-warning signal before the formal stop triggers.

Thesis Scorecard

SaaS Application Displacement

• WDAY ($123.83): 4-source convergence; Duffield 10%-holder sold 4× in 3 weeks (-51% stake each = $53.8M); CFO + 2 Presidents distributing; RegSHO 0.7053 elevated, 10d max 0.818; SBC 50% of rev → FCF 64% is engineering, GAAP op_margin 11% is the real number. Entry $124 / stop $136 / target $111.45.
• DOCU ($45.74): highest 4-signal internal convergence in thesis universe (RegSHO 0.756 anomaly + slope +0.182 + SI +22.1% MoM + pc_vol 2.24); CEO/CRO/GC/Dir all sold Apr 1-3 in same $46.83-48.15 band. Entry $45.75 / stop $48.50 / target $41.17.
• TEAM ($66.94): direct Apr 30 earnings test of NRR/per-seat thesis; TWLO same session amplifies vol. Entry $67 / stop $71.50 / target $60.25; 5-day time-stop.
• ZM ($88.02): tri-level C-suite liquidation in 10-day window (CEO Yuan -70% / COO -87% / CFO -22%). Entry $88 / stop $93.50 / target $79.22.
• CRM, ASAN, DOMO, MNDY, HUBS, PATH: monitor only — CRM SI data quality-flagged (+255% MoM likely artifact); ASAN squeeze risk (DTC 5.25, SPF 32%); PATH SI unwinding -7.25% MoM.

AI Hardware & Power

• NVDA ($201.68): 4-signal convergence (market/innovation/macro/catalyst). ICLR 21d + Apr 29 hyperscaler tone. Entry $201 / stop $191 / target $221. RenTech trim partial offset.
• MU ($455.07): rev +196% YoY, op_margin 68%, SBC 2.5% — cleanest growth name in dataset. HBM ↔ Blackwell. Entry $452 / stop $432 / target $500.
• VRT ($307.34): ONLY positive deferred-rev in entire fundamentals dataset (+36% QoQ). At 52wk high. Entry $307 / stop $291 / target $338.
• AMD ($278.39): May 5 earnings event_step; rev +36% YoY, FCF 47%. Entry $278 / stop $264 / target $306; size halve if hyperscaler Apr 29 negative.
• AVGO/TSM/SMH/CEG/SMCI: watchlist; SMH ETF expression preferred to single-name pre-Apr-29 binary.

Cybersecurity Consolidation

• FTNT ($81.84): May 6 earnings (FOMC overlap); EPS $0.53 consensus. Entry $82 / stop $77.75 / target $90.02; event_step.
• ZS ($134.68): ZS-OpenAI zero-trust partnership (Apr 17) validates platform consolidation; pullback entry. Entry $131 / stop $127.95 / target $148.15.
• CRWD ($423.95): NO entry until post-MSFT Apr 29; only net-bearish StockTwits name (-0.13) during sector rally — community pricing in headwind not yet in headlines.
• PANW ($167.85): scale on hold above $162; await break $176 confirmation.
• MSFT Apr 29 Defender language is the single highest-impact binary for the entire thesis.

Infrastructure Software

• PLTR ($146.39): sole bullish convergence in thesis universe (catalyst + fundamentals); rev +63% YoY, opmargin 33%, FCF 113%. May 4 earnings eventstep. Entry $147 / stop $132 / target $172. 14% model hit-rate caps conviction.
• DDOG ($126.61): wait for $123-126 pullback after +15.3% 5d surge. Entry $125 / stop $116 / target $142.
• MDB ($263.38): momentum-only thesis; no Q1 fundamentals available. Entry $265 / stop $245 / target $292.
• NET ($200.99): downgraded to neutral — CEO Prince -100% delta_owned at $211.94, stock now -5.4% below his exit; CFO + Co-Chair also liquidating ($51.8M aggregate); thesis on hold pending 90d insider overhang clear.
• SNOW: SBC 99% of revenue makes reported FCF 29% pure add-back; no entry.

Energy Security

• XOM ($146.44): May 1 earnings reflect Q1's $70+ WTI environment, not the April spot collapse. Hormuz 72.7% disruption vs Apr-30 normalization 2.85% = sustained tail. Entry $146.50 / stop $138 / target $158; event_step.
• CVX ($183.99): integrated buffer (LNG + chemicals) preferred in demand-fear regime. Entry $184 / stop $174 / target $197.
• OXY ($53.79): -5.4% single-day dislocation into Berkshire ~$52 cost basis. Limit buy $52 scaling to $51.10. Stop $48 / target $58.50.
• COP/XLE/SLB: wait for crude to stabilize >$65 or geopolitical confirmation.

Macro & Credit Regime

• TLT ($87.07): SHORT into FOMC May 6 + Warsh confirmation. Entry $87.50 / stop $91 / target $83.50. Three-source: Warsh 94.4% + CPI re-accel + FOMC tone trending hawkish. Asymmetric to window high 90.16.
• HYG ($80.65): SHORT credit complacency. Exact window_high. OAS 2.84% near cycle tights vs 26% recession prob. Entry $80.65 / stop $82 / target $77.60.
• SPY put hedge ($710.14): tactical only at exact 160d window high; F&G 66 + safe-haven 99.4% bi-modal. Define-risk via May 16 700/685 put spread. Stop above 730 close / target $675.
• VIX/IWM/XLF/XLK/IGV: monitor — VIX 17.48 declining into peak earnings density is favorable vol-buying setup.

FDA Pharmaceutical Catalysts

• LLY ($927.03): RenTech AND DE Shaw both initiated new positions Q4 2025 — only dual-quant initiation in dataset. ASCO 52d. Entry $928 / stop $880.68 / target $1019.73.
• AMGN ($355.30): RegSHO double-anomaly (0.798 + deviation +0.304) — most extreme biotech short escalation in universe; no news catalyst = institutional positioning. Entry $354 / stop $373.07 / target $319.77.
• MRK, VRTX, XBI, PFE, REGN, ABBV, GILD, BMY: monitor; calendar desert (zero core PDUFAs in 60d); ASCO is the next material event-cluster.

Quantum & Emerging Tech

• GOOG ($339.40): Apr 29 earnings — Willow chip roadmap + AI capex compound catalyst. Entry $339.40 / stop $322.43 / target $356.37; event_step.
• IONQ ($46.09): May 6 earnings (no EPS estimate available); FOMC hold same day. Entry $47 / stop $41.48 / target $50.70; 0.5% sizing — pure-play after +56% 5d surge with no fundamental anchor.
• IBM, RGTI, QBTS, MSFT, NVDA: sector coattail; pure-plays (RGTI +33%, QBTS +48% 5d) speculative — wait for $17.83-19.52 / $19.52 supports.

Cross-Thesis Convergences

WDAY (bearish, 4 sources) — short: RegSHO 0.7053 elevated, 10d max 0.818, historical pcratio 3.89 Apr 15; fundamentals: SBC 50% of rev, GAAP opmargin 11% real vs FCF 64% engineered; insider: 8 sellers 90d, $56.6M net sell, Duffield 10%-holder -107.5K shares 4× in 3wk (~$53.8M, -51% stake each), CFO + 2 Presidents distributing; macro: named in aicapabilityleap (p=0.40) AND warshfedchair_confirmation (p=0.944) tail tickers.

DOCU (bearish, 3 sources) — short: 4-signal internal (RegSHO 0.756 > 0.75 anomaly, slope +0.182, SI +22.1% MoM, pc_vol 2.24); insider: CEO Thygesen + CRO Hansen + GC Shaughnessy + Dir Wilderotter all sold Apr 1-3 at $46.83-48.15 (same trade-date range, $2.24M aggregate); macro: DOCU/ZM/WDAY/NET app-SaaS distribution cluster, Warsh 94.4% long-duration headwind.

CRM (bearish, 3 sources) — short: simomchange +2.5526 (FLAGGED as likely yfinance artifact, +255% MoM), RegSHO 0.7015 elevated+rising delta +0.190; sentiment: Wikipedia views +91.6% velocity (49 vs 25.6 baseline) — highest surge in SaaS universe; macro: in aicapabilityleap and Warsh tail tickers. Excluded from trade table due to SI quality flag.

DOMO (bearish, 3 sources) — market: +11.1% single-day Apr 18 anomalycandidate; fundamentals: deferred rev -5% QoQ on flat rev (-0.5% YoY), SBC 51%, opmargin -9%; sentiment: StockTwits +0.33 (33% bull / 0% bear) with zero fundamental support = textbook distribution.

MSFT (mixed, 3 sources) — catalyst: Apr 29 earnings, EPS $4.07; fundamentals: deferred rev -13% QoQ despite rev +17% YoY (Azure/M365 mix shift signal); insider: 2 adding vs 2 trimming, net -1.9M shares, Tiger Global $3.7B largest. Both longinfrasw and short_saas thesis-aligned — single most consequential ticker for portfolio Apr 29.

ASAN, NET, ZM, PLTR, CHGG, BBAI: 2-source convergences — see thesis deep dives.

Catalyst Watch

14d window: - Apr 29 (5d) — MEGA-CLUSTER 11+ reporters: MSFT $4.07, GOOG $2.64, AMZN $1.61, META $6.71 + FVRR $0.42, ABBV, KLAC, QCOM, EQIX, GD, REGN. AI capex tone is the single highest-impact session. - Apr 30 (6d) — AAPL $1.91; TWLO $0.56 + TEAM $0.09 (per-seat displacement test); GTLS (Oracle helium probe). - May 1 (7d) — XOM $1.21, CVX $1.09, LIN $4.27 (helium). - May 4 (10d) — PLTR $0.22 (model blind-spot 14% hit rate; gap_driven query active). - May 5 (11d) — AMD $1.04, PFE $0.74, SHOP $0.22, ANET $0.72. - May 6 (12d) — FOMC hold 99.45% implied + Warsh confirmation watch + ARM $0.37, FTNT $0.53, NRG $1.62, IONQ no estimate, ALB. Macro binary embedded in dense sector session.

30d window: May 31 (37d) CELC VIKTORIA-1 PIK3CA cohort readout — Oracle probe.

60d window: Jun 14-17 ASCO 2026 (51-54d) — pharma event-cluster anchor; Jun 17 (54d) FOMC + Jun 17 CELC Gedatolisib PDUFA action-window opens.

Action windows: Apr 29 + May 6 are the two binaries to size around — vol compression into both is favorable for define-risk option structures.

Probability Dislocations

Empty from integrator this cycle (probability_dislocations: []). Two informal candidates to track manually:

1. Hormuz disruption: signals composite 72.7% vs Polymarket Apr-30 normalization 2.85% — implies model sees ~25× higher disruption persistence than market; XOM/CVX/OXY long thesis derives directly from this dislocation if it is real; resolves by Apr 30 (6d).

2. HY OAS 2.84% near cycle tights vs 26% recession probability — historically these levels precede spread widening within 30-60d in 4 of 5 comparable episodes (per macro reader). HYG short trade monetizes this dislocation.

Both dislocations are unverified by Polymarket-derived edge calculation because fda_pipeline.json (480KB) and full prediction-market scrape exceeded reader limits this cycle.

Risk Factors

1. Apr 29 hyperscaler quad invalidates 7 of 10 trades simultaneously. MSFT/GOOG/AMZN/META beating on AI capex re-rates the entire long-hardware leg up but triggers a SaaS short-covering acceleration (IGV already +8.6% 5d) that stops out WDAY/DOCU/TEAM. Concentration risk: size correlated trades as ONE position not seven.

2. IGV +8.6% 5d covering rally not yet exhausted. If IGV breaks above $89.33 (5pct resistance from $85.08 close), forced covering across crowded SaaS shorts overrides fundamental signals across the basket. Wait for IGV daily RSI > 70 with failure below 88 before adding.

3. Market data 4-6 trading days stale (digest latest_day 2026-04-18); stimpy-3 pipeline likely not running. All entry/stop/target levels reference Apr 18 closes — actual current prices may differ materially. Verify before any execution. This is the single most urgent operational gap.

4. Warsh confirmation reversal (6% Polymarket). Failed nomination → dovish replacement → TLT short and HYG short both reverse simultaneously; long-duration SaaS rallies, breaking short_saas multiple-compression thesis.

5. GDP nowcast surprises upward in Q1 final. Soft landing confirmation vs current 0.5% annualized stall reading — credit complacency justified, recession-probability shorts (HYG, IWM, SPY hedge) all reverse together.

Sentiment & Positioning

Cross-current backdrop: F&G at 66 (greed) simultaneously with safe-haven demand at 99.4% (extreme greed) = bi-modal institutional uncertainty, not consensus. VIX declining 19.38 → 17.48 over 5d into peak earnings density (Apr 29 + FOMC May 6) = vol-buying setup. WCLD (cloud SaaS ETF) put/call 47.0 (1 call vs 47 puts) = institutional sector-level de-risking confirms WDAY/DOCU/TEAM short thesis at basket level. CRM Wikipedia views +91.6% velocity = retail-driven runup pre-Q1 earnings = fade signal post-Apr-29 hyperscaler. CRWD only net-bearish StockTwits ticker (-0.13, 37% bear vs 23% bull) during a broad security rally = sector-specific community concern preceding fundamentals (typical 2-4 week lead). Macro hedge surge (SPY uvr 19.6× + QQQ uvr 12.8× simultaneous) signals portfolio-level tail hedging across the board.

Fundamental Outliers

• VRT: deferred rev +36% QoQ — the ONLY positive deferred-rev entry in the entire fundamentals dataset; rev +30% YoY. AI thermal management bookings genuinely accelerating while every SaaS name shows deferred declining.
• MU: rev +196% YoY at op_margin 68%, SBC 2.5%. Cleanest growth quality in universe — flagged explicitly.
• PLTR: rev +63% YoY, op_margin 33%, FCF margin 113%. Best fundamental quality in infra-sw long thesis.
• ANET: rev +27% YoY, op_margin 42%, FCF 131%. Highest FCF margin in dataset; excluded from trades only because absent from price-grounding file (recommend universe expansion).
• SNOW: SBC 99% of revenue — reported FCF 29% is essentially pure SBC add-back with zero true cash generation; op_margin -27%. Worst earnings-quality outlier in software universe.
• MSFT: deferred rev -13% QoQ despite revenue +17% YoY — Azure/M365 bookings mix shift signal; not yet narrative.
• ASAN: SBC 82% of rev, opmargin -35%, rev growth slowing to 9%; fundamentallyweak.
• CHGG: rev -43% YoY; deferred rev -7.5% QoQ; SBC 32% of shrinking base; op_margin -22% — AI tutoring disruption fully manifest.
• BBAI: rev -20% YoY AND deferred rev -21% QoQ — double-confirmation of forward-bookings collapse; op_margin -66%.
• DOMO: deferred rev -5% QoQ on flat rev (-0.5% YoY), SBC 51%, op_margin -9%.

Insider & Institutional Flow

App-SaaS C-suite distribution cluster (statistically improbable under routine 10b5-1 hypothesis alone): - WDAY: Duffield 10%-holder sold 107.5K shares 4× in 3 weeks (~$53.8M, -51% stake each); CFO + 2 Presidents selling; 8 sellers 90d, $56.6M net sell. - DOCU: CEO Thygesen + CRO Hansen + GC Shaughnessy + Dir Wilderotter all sold Apr 1-3 in $46.83-48.15 range = coordinated full-exec flush ($2.24M aggregate at known resistance). - NET: CEO Prince sold 156,493 shares at $211.94 with -100% deltaowned (effective full liquidation via S+OE); CFO Seifert -$2M; Co-Chair Zatlyn -$15.8M; $51.8M 90d aggregate. Stock now -5.4% below CEO exit — primary thesis ticker invalidated for longinfra_sw thesis. - ZM: CEO Yuan -70% directly-held stake ($6.9M), COO Bawa -87% ($1.1M), CFO Chang -22% ($690K) in same 10-day window.

Aggregate: >$120M C-suite net sells across 4 thesis tickers in same month = the dominant insider signal of the cycle.

Institutional accumulation: - LLY: RenTech AND DE Shaw both initiated new positions Q4 2025 13F simultaneously — single highest-conviction quant institutional signal in dataset (4-month staleness caveat). - NVDA: RenTech trimming -4.6M shares vs DE Shaw adding — quant-fund disagreement, slightly net negative. - MSFT: 2 adding vs 2 trimming, net -1.9M shares (marginal); Tiger Global $3.7B largest. - AMGN: RegSHO double-anomaly (ratio 0.798 > 0.75 AND +0.304 deviation from 10d mean) without news catalyst = unidentified institutional short pressure. - No insider buy/sell data emitted for energy thesis tickers (XOM/CVX/OXY/COP) this cycle — ambiguous (data gap vs absence of activity).

Research & Innovation

Forward-looking R&D signals (under-priced source, ~ 2-6 week lead-time on positioning):

• ICLR 2026 in 21 days — active cs.CL SOTA claim flagged as fast_deployment category; NVDA direct beneficiary, ANET/AMD adjacent. Conference proximity historically drives pre-event accumulation in semis. Add to NVDA conviction.
• Quantum (quant-ph) SOTA claim Apr 23: tensor network belief propagation paper architecture-relevant to IonQ trapped-ion stack. Photonic-switch paper (universal quantum network routing) directly relevant to IONQ photonic interconnect roadmap.
• ASCO 2026 (Jun 14-17, 52d) — pharma event-cluster anchor. Zero corporate medRxiv preprints from PFE/MRK/LLY/REGN in 30d window despite ASCO 52d out: either (a) all four firms holding data under embargo (bullish surprise risk to LLY long) or (b) ingestor failure. Validate urgently — primary early-signal channel.
• Quantum corporate preprints empty across IONQ/RGTI/QBTS/IBM/GOOG 30d — likely ingestor failure not absence; treat innovation silence as null-signal not confirmation. Critical for sizing IONQ May 6 reaction function.
• bioRxiv/medRxiv corporate preprints completely empty in 30-day window across full coverage — high-priority ingestor repair item; this is the main forward catalyst-leading-indicator channel and it is not running.

Citation breakouts / arXiv velocity spikes: none surfaced this cycle (innovation reader limited this cycle); add to backlog for next manifest.

Where This Analysis Could Be Wrong

1. Apr 29 hyperscaler quad beats AI capex broadly → entire shortsaas thesis reverses single-session; IGV breaks 90+; WDAY/DOCU/TEAM/ZM all stop out simultaneously while longhardware book gaps in our favor — net P&L still likely negative because SaaS shorts are sized larger in aggregate (4.5% notional) than NVDA+MU+VRT+AMD (5%) net of correlation. Falsifier: MSFT Azure rev +20%+ YoY AND SaaS displacement commentary either absent or rejected on Apr 29 call.

2. CRM SI data (+255% MoM) is real, not artifact → triggers actual squeeze; IGV-led basket short-cover cascades; the short_saas integrator convergence we relied on becomes a reverse-indicator. Falsifier: verify with Bloomberg/Markit raw SI data before next cycle; if confirmed, CRM gets added to short basket; if rejected, broader integrator confidence calibration improves.

3. WDAY/DOCU/ZM/NET insider sales are pre-scheduled 10b5-1 mechanics + tax/estate planning, not information-driven → the cluster signal we treat as the strongest insider signal of the cycle is noise. Falsifier: S+OE plan filing dates pre-date the FY26 budget cycle and are routine; we lack this data point for any of the four tickers — high-priority ingest item.

4. Hormuz 72.7% disruption probability is a model artifact from over-weighting headlines → the energy long thesis (XOM/CVX/OXY) loses its asymmetric kicker; XOM rallies on Q1 earnings beat but fails to extend. Falsifier: Polymarket Hormuz contracts diverge from our composite by >40pp by Apr 30; or our composite halves within 7 days without an underlying news driver.

5. Market data being 4-6 trading days stale means the 'window high' setups (SPY 710.14, HYG 80.65, IWM 275.78) are no longer valid — any of those tickers may have already broken out or down. Sizing decisions made on stale technical levels may be at materially wrong prices. Falsifier: restart stimpy-3 pipeline; re-pull current Apr 24 closes; if any of SPY/HYG/IWM/TLT have moved >2% from grounding values, re-derive entries before execution.

Oracle Attention — User-Surfaced Tickers

Oracle-surfaced tickers (probe watchlist): - GTLS — Chart Industries Q1 2026 earnings Apr 30 (6d). Helium picks-and-shovels (majors/picksandshovels affinity). Blue Spruce contract + cryogenic order backlog are the watch-metrics. Not in core universe; magnitude medium. Recommended probe: add GTLS to next-cycle market ingest manifest; do not trade this cycle (no price grounding). - CELC — VIKTORIA-1 PIK3CA-mutant cohort readout est May 31 (37d); Gedatolisib PDUFA Jul 17 (84d, action window opens ~Jun 17 alongside June FOMC). Less-priced-in than PDUFA per Oracle. Oracle query_count=2. Competitors: NVS (alpelisib), RHHBY (inavolisib), AZN (capivasertib). Recommended probe: add CELC to Polymarket tracking for approval probability; monitor for ASCO abstract or medRxiv preprint as actionable trigger. - MAZE — $150M raise completed Apr 22; Phase 1/2 kidney/metabolic biotech; no near-term PDUFA. Status: observe only; no trade trigger.

Open thesis candidates: none surfaced from candidatestresshits this cycle (searchmanifest thesiscandidate_tests not directly inspected — quality flag from integrator).

Generated by synthesis-v2 DAG on 2026-04-24