Requested 2026-09-19 08:51:24.623028 UTC
· nvidia/nemotron-3-ultra-550b-a55b:free
· window 2026Q1–2026Q2
· finished 2026-09-19 09:10:36.060082 UTC
EBERT CAPITAL MANAGEMENT INC. — 13F FORENSICS: 2026Q1 → 2026Q2
Mandate: Not stated. Inferred style: Concentrated US Growth / AI Thematic with Tactical Overlay (Cash mgmt, Option Income, Gold Hedge). Long-only reportable sleeve only.
BOTTOM LINE (5 BULLETS)
- Massive Cash Redeployment — BIL (short Treasury ETF) slashed −5.2 pp (20.2% → 15.1%), funding concentrated AI adds and new hedge/overlay sleeves. Verified (weights).
- Active Beta Reduction — QQQ shares trimmed (−581) despite weight rising +2.1 pp to 16.8% (price-driven). Manager is selling broad Nasdaq beta into strength to fund single-name AI. Inferred (Δshares vs Δweight divergence).
- AI Concentration Hardening — NVDA +686 shrs (20.2% wt), AVGO +255, SMH NEW 0.3%, GOOGL +168, META +729. Semis + Mega-cap AI now ~39% of sleeve (incl. QQQ overlap). Verified (weights, Δshares).
- Distinct "Adding into Weakness" Signals — NFLX +2,603 shrs while weight −0.6 pp; MSFT +304 shrs while weight −0.3 pp. High-conviction dip-buying contradicted by price lag. Verified (Δshares vs Δweight).
- New Hedge/Income Layer — GLD (0.5%), SPCX covered-call ETF (0.6%), SMH (0.3%) initiated. BIL drawdown + option overlay + gold = explicit downside/volatility hedge construction. Verified (New positions).
1️⃣ PORTFOLIO STRUCTURE QoQ (VERIFIED FROM WEIGHTS)
| Metric |
2026Q1 (Baseline) |
2026Q2 |
Δ |
Note |
| Active Positions |
26 |
28 |
+2 |
+3 New (GLD, SMH, SPCX), −1 Exit (SHOP) |
| Top-5 Weight |
65.9% |
64.7% |
−1.2 pp |
NVDA, BIL, QQQ, GOOGL, MSFT |
| Top-10 Weight |
~82.0% |
~82.0% |
~Flat |
Extreme concentration persistent |
| HHI (Σw²) |
0.105 |
0.099 |
−0.006 |
Marginally diversified; still "high conviction" |
| Cash Proxy (BIL) |
20.2% |
15.1% |
−5.2 pp |
Largest single portfolio move |
| ETF/Overlay Sleeve |
41.4% |
35.5% |
−5.9 pp |
BIL/QQQ/SPY/FDN/IGV trimmed; SPCX/GLD/SMH added |
Formula: HHI = Σ(weight_pct/100)². Top-N = sum of largest N weights. ETF Sleeve = sum of all ETF weights (BIL, QQQ, SPY, SPYG, FDN, IGV, GLD, SMH, SPCX).
2️⃣ TOP-10 2026Q2 + MATERIAL CHANGES (VERIFIED WEIGHTS & ΔSHARES)
| Rank |
Ticker |
2026Q2 Wt |
Δ Wt (pp) |
Q2 Δ Shares |
Action |
Flow Signal |
| 1 |
NVDA |
20.21% |
+1.16 |
+686 |
ADD |
Core AI, adding into strength |
| 2 |
QQQ |
16.75% |
+2.12 |
−581 |
TRIM |
Selling into strength (beta reduction) |
| 3 |
BIL |
15.06% |
−5.18 |
−41,135 |
TRIM |
Massive cash deployment |
| 4 |
GOOGL |
7.23% |
+0.95 |
+168 |
ADD |
Core AI/Cloud, adding into strength |
| 5 |
MSFT |
5.40% |
−0.32 |
+304 |
ADD |
Adding into weakness (wt ↓, shrs ↑) |
| 6 |
AMZN |
5.02% |
+0.28 |
+142 |
ADD |
Core Cloud/Retail |
| 7 |
FICO |
3.66% |
+0.26 |
+143 |
ADD |
Fintech/Data core |
| 8 |
COST |
3.67% |
−0.60 |
−19 |
TRIM |
Consumer staple trim |
| 9 |
V |
2.31% |
+0.15 |
+188 |
ADD |
Payments core |
| 10 |
LLY |
2.45% |
+0.49 |
+87 |
ADD |
Healthcare core |
Material New / Exits / ±25% Scaled (Verified)
| Category | Ticker | Q1 Wt | Q2 Wt | Δ Shares | Verdict |
|---|---|---|---|---|---|
| NEW | SPCX | 0.00% | 0.61% | +3,784 | Covered-call income overlay |
| NEW | GLD | 0.00% | 0.46% | +1,317 | Gold hedge |
| NEW | SMH | 0.00% | 0.30% | +493 | Semi basket (broadens NVDA/AVGO) |
| EXIT | SHOP | 0.45% | 0.00% | −3,679 | Full exit |
| ≥+25% Wt | NFLX | 2.45% | 1.84% | +2,603 | Weight −25%, Shares +100%+ → Adding into severe weakness |
| ≥+25% Wt | AVGO | 1.18% | 1.41% | +255 | Weight +20% (near threshold) |
| ≥−25% Wt | BIL | 20.24% | 15.06% | −41,135 | Weight −26% (cash deployment) |
| ≥−25% Wt | IGV | 2.24% | 2.10% | −2,698 | Weight −6% (software ETF trim) |
| ≥−25% Wt | NOW | 0.45% | 0.37% | −210 | Weight −17% (software trim) |
Note: "±25% Scaled" = |ΔWeight|/Q1_Weight ≥ 25%. NFLX is the only position with large share increase + large weight drop — textbook "adding into weakness".
3️⃣ CONVICTION RANK — Q2 ΔSHARES (RAW, UNSCALED)
No share counts or prices → dollar conviction LOW CONFIDENCE. Ranked by raw Δshares (decision magnitude).
| Conviction Tier |
Ticker |
Δ Shares |
Direction |
vs Weight Δ |
vs Price (Inferred) |
| 1. MAX REDUCE |
BIL |
−41,135 |
SELL |
Wt −5.2 pp |
Price stable → Pure flow |
| 2. MAX ADD (New) |
SPCX |
+3,784 |
BUY |
Wt +0.6 pp |
New position |
| 3. MAX ADD (Weak) |
NFLX |
+2,603 |
BUY |
Wt −0.6 pp |
Adding into weakness |
| 4. EXIT |
SHOP |
−3,679 |
SELL |
Wt −0.45 pp |
Full liquidation |
| 5. REDUCE |
IGV |
−2,698 |
SELL |
Wt −0.14 pp |
Software beta trim |
| 6. ADD (New) |
GLD |
+1,317 |
BUY |
Wt +0.46 pp |
New hedge |
| 7. ADD |
META |
+729 |
BUY |
Wt +0.27 pp |
Adding into strength |
| 8. ADD |
NVDA |
+686 |
BUY |
Wt +1.16 pp |
Adding into strength |
| 9. REDUCE |
QQQ |
−581 |
SELL |
Wt +2.12 pp |
Selling into strength |
| 10. ADD |
MSFT |
+304 |
BUY |
Wt −0.32 pp |
Adding into weakness |
| … |
AVGO, MA, SMH, V, FICO, GOOGL, AMZN, LLY, SPGI, AAPL, SPY, FDN, NOW, MSCI, COST |
< |
Δ300 |
|
|
Cross-tab insight:
- Adding into Strength: NVDA, GOOGL, META, AVGO, V, LLY, AMZN (Δshrs >0, Δwt >0).
- Adding into Weakness: NFLX, MSFT, FICO (Δshrs >0, Δwt ≤0). Speculative: requires price confirmation.
- Selling into Strength: QQQ, AAPL (Δshrs <0, Δwt >0).
- Selling into Weakness: BIL (price stable), IGV, NOW, SHOP, COST, MSCI, SPY (Δshrs <0, Δwt <0).
4️⃣ SUB-SECTOR ROTATIONS → MACRO ASSUMPTIONS (INFERRED)
A. Cash & Rates Overlay: BIL → GLD + SPCX + Equities
- Flow: BIL −$41k shrs (est. −$3.7B notional); GLD +1,317 shrs; SPCX +3,784 shrs.
- Macro Assumption: Fed cutting cycle imminent / "Higher for Longer" over. Cash drag unacceptable. Gold = fiscal debasement / geopolitical hedge. SPCX (covered call) = expectation of range-bound / high-vol equity markets where option premium > upside participation.
- Confidence: High (Verified flows, clear macro mapping).
B. Broad Tech Beta → Concentrated AI Winners
- Flow: QQQ −581 shrs, IGV −2,698, FDN −156, NOW −210, MSCI −57 → NVDA +686, AVGO +255, SMH NEW, GOOGL +168, META +729.
- Macro Assumption: AI CapEx cycle broadening (AVGO networking, SMH breadth) but alpha only in oligopolists. Broad software (IGV, NOW) multiple compression anticipated. Passive beta (QQQ) used as liquidity reservoir.
- Confidence: High (Verified flows, consistent across 5+ tickers).
C. Fintech / Financial Infrastructure Upgrade
- Flow: V +188, MA +254, SPGI +174, FICO +143 → MSCI −57.
- Macro Assumption: Payment volumes & pricing power durable (V/MA); Credit/analytics demand structural (FICO/SPGI). MSCI trim = index-license saturation / ETF flow sensitivity.
- Confidence: Medium (Small Δshares, but consistent direction).
D. Consumer Selectivity: SHOP Exit / COST Trim / NFLX Dip-Buy / AMZN Add
- Flow: SHOP EXIT (−3,679), COST −19, NFLX +2,603 (wt −25%), AMZN +142.
- Macro Assumption: SHOP platform risk (competition, take-rate pressure). COST valuation full. NFLX advertising tier + paid sharing inflection → buying aggressive dip. AMZN AWS re-acceleration + retail margin.
- Confidence: Medium-High (SHOP exit decisive; NFLX divergence stark).
E. Healthcare: LLY Add / JHML Hold
- Flow: LLY +87 shrs (wt +0.5 pp). JHML unchanged.
- Macro Assumption: GLP-1 monopoly durability (LLY). JHML (Janus Henderson Long-term Care ETF?) held as demographic hedge.
- Confidence: Low (Single name, small flow).
5️⃣ CAP / LIQUIDITY / LEVERAGE / GEOGRAPHY TILTS (VERIFIED / INFERRED)
| Dimension |
2026Q1 |
2026Q2 |
Drift |
Assessment |
| Market Cap |
Mega/Large only |
Mega/Large only |
None |
100% >$50B; zero small/mid direct. Verified |
| Liquidity |
Extreme (ETFs + Mega) |
Extreme (ETFs + Mega) |
None |
All positions screen highly liquid. Verified |
| Leverage |
None visible |
SPCX (Covered Call) |
New |
Implicit leverage via option overlay; no balance-sheet leverage visible in 13F. Inferred |
| Geography |
100% US-listed |
100% US-listed |
None |
No ADRs, foreign ords, EM. Verified |
| Sector |
Tech 45%, Cash 20%, Fin 10%, Cons 7%, HC 2% |
Tech 50%, Cash 15%, Fin 12%, Cons 10%, HC 3%, Hedge 1% |
Tech ↑, Cash ↓, Hedge ↑ |
Inferred from buckets |
6️⃣ STRATEGY IMPLIED BY FLOWS (INFERRED — 2 QUARTERS ONLY)
| Pillar |
Evidence |
Confidence |
| Concentrated AI/Thematic Core |
Top 3 = NVDA (20%), QQQ (17%), GOOGL (7%) = 44%. NVDA/AVGO/SMH/GOOGL/META adds. |
High |
| Tactical Cash Management |
BIL 20% → 15% in one quarter; not a strategic reserve, but dry powder. |
High |
| Active Beta Management |
QQQ shares sold while weight rises → explicit beta reduction vs passive. |
High |
| Overlay Hedge Construction |
GLD + SPCX new = 1.1% dedicated hedge sleeve (was 0%). |
High |
| Conviction Dip-Buying |
NFLX (+2.6k shrs, wt −25%), MSFT (+304, wt −5%). |
Medium |
| Software De-rating |
IGV, NOW, MSCI, FDN all trimmed. FICO/SPGI only adds. |
Medium |
Contradictions / Open Questions
1. QQQ Paradox: 2nd largest position (16.8%) but only position with Δshares < 0 AND Δweight > +2 pp. Is QQQ a "core" holding or a liquidity sleeve for funding NVDA/AVGO/META? Flows say liquidity sleeve; weight says core.
2. NFLX Conviction vs Price: Largest share increase in portfolio (+2,603) but weight collapsed. Is this high-conviction value investing or catching a falling knife? Next filing decides.
3. SPCX Sizing: 0.6% weight but +3,784 shares (3rd largest raw Δshares). If SPCX ~$22/sh, that’s only ~$83k notional — tiny. If ~$100/sh, ~$378k. Without price, conviction scaling is unknowable. Low Confidence.
4. Mandate Drift: 2026Q1 had 20% cash (BIL). 2026Q2 has 15% cash + 1% hedge + 5% new risk. One-quarter rebalance ≠ strategy. Could be single rebalance event (e.g., LP capital call, risk limit change).
7️⃣ 3–5 FALSIFIABLE PREDICTIONS FOR 2026Q3 FILING
| # |
Prediction |
If Wrong (Next Filing Shows…) |
| 1 |
BIL weight < 10% (continued deployment to ~5-8%) |
BIL weight ≥ 15% → cash deployment paused/reversed |
| 2 |
QQQ shares decline further (active beta reduction continues) |
QQQ shares flat/up → QQQ re-promoted to core |
| 3 |
SPCX weight > 1.0% (overlay scaling) |
SPCX weight < 0.4% or exited → trial hedge abandoned |
| 4 |
NFLX shares increase (doubling down on dip) |
NFLX shares flat/down → dip-buy abandoned |
| 5 |
SHOP remains exited (no re-entry) |
SHOP reappears → exit was tactical/tax-loss |
8️⃣ DATA LIMITATIONS & EPISTEMIC FLAGS
| Limitation |
Impact |
Mitigation |
| Only 2 quarters (Q1 baseline, Q2 flow) |
Small sample — may reflect one rebalance, not strategy. |
Label all strategy claims Inferred; require 4+ quarters for "Verified Strategy". |
| No share counts, prices, AUM |
Cannot compute: $ turnover, bridge (flow vs price), dollar conviction, % position scaling. |
Use raw Δshares for flow direction only; mark all $ metrics Not Computable / Low Confidence. |
| 2026 dates = future |
Prices unknowable; corporate actions post-2024 unverifiable. |
Flag all price-dependent inferences Speculative. |
| 13F blind spots |
No shorts, OTC, derivatives (ex reportable converts), cash (ex BIL), private, fixed income. |
Call it "Reportable Sleeve" — not total portfolio. |
| 45-day lag |
Q2 filing reflects June 30; analysis dated ~Aug 15. |
Stale for tactical positions (SPCX, GLD). |
| Delta_shares in Q1 |
Left-censored — mechanically "NEW". Ignored per rules. |
Only Q2 Δshares used for conviction. |
9️⃣ CONTRADICTIONS SUMMARY (FLOWS VS IMPLIED CONVICTIONS)
| Implied View |
Flow Evidence |
Contradiction? |
| "Long-term AI conviction" |
NVDA +686, AVGO +255, SMH NEW |
Consistent |
| "Software overweight" |
IGV −2,698, NOW −210, MSCI −57, FDN −156 |
Contradicted — software beta reduced |
| "Cash conservative" |
BIL 20% → 15% in 1 quarter |
Contradicted — aggressive deployment |
| "Passive core" |
QQQ −581 shrs (sold) |
Contradicted — active beta management |
| "NFLX high conviction" |
+2,603 shrs (largest add) |
Supported by flow, contradicted by price |
| "Fintech core" |
V, MA, SPGI, FICO all adds |
Consistent |
FINAL NOTE: This sleeve behaves like a long-biased hedge fund masquerading as long-only: high concentration (NVDA 20%), active beta trading (QQQ), cash as tactical tool (BIL), and new option/Gold overlays (SPCX, GLD). The next filing (2026Q3) will reveal whether the BIL→Risk rotation completes (BIL <10%), whether NFLX dip-buy persists, and whether SPCX scales — the three signals that separate strategy from single-quarter rebalance.