Three-Stock Tech Sleeve — Execution Thesis (MSFT · NFLX · PLTR) #

Date: 2026-06-26 · Budget: €3,000 · Broker: DEGIRO (manual, whole-share, USD) · FX: EUR/USD ≈ 1.17 → ~$3,510 deployable Supersedes: the 4-stock draft (INTC dropped 2026-06-25 — negative margin of safety, traded +37% above analyst PT; see §6).

Governance (read first). Standalone tactical sleeve, user-directed, departing from charter §6/§7 (single-name ban; NFLX is the only logged §8 exception). MSFT & NFLX are quality names you already own inside IS3Q (§5); PLTR is deliberately excluded by the quality factor and is a conscious anti-factor growth bet. Track separately from the legacy Graham book and the evidence-corpus tilt. This sleeve is β≈1.29, σ≈28% — three correlated US-tech names, weak internal diversification. It is an aggressive conviction overlay, not core. Per shr-020/041/042, re-verify live price, share count and EUR/USD in DEGIRO before any fill.

1. Allocation — the €3,000 #

How sizing was derived (full reasoning in §4): inverse-volatility base (DeMiguel 2009 — don't optimize 3 noisy return forecasts; Maillard-Roncalli risk parity) → quality/low-beta/Graham-MoS tilt (QMJ, BAB, Graham growth-addendum) → freed INTC budget flowed to the lowest-vol, positive-MoS anchor (MSFT). Net effect vs the 4-stock plan: β1.42/σ32% → β1.29/σ28%.

2. Per-name execution #

MSFT — $352.83 · anchor (52.8%) #

• Why: highest quality (ROE 33%, ROIC 26%, AAA, net cash ~$34B, op margin 47%), lowest beta (1.10), and the only genuine Graham margin of safety: implied growth from fwd P/E 18.2x = 4.9% vs ~18–29% delivered; Graham IV @10%g ≈ $467 = +32% MoS. Trades below even the most bearish analyst PT ($400).
• Targets (compounder — re-assess, don't auto-sell): T1 $465 (+32%, fair value) · T2 $555 (+57%, analyst mean — trim ⅓ if growth also slowing) · T3 $630 (+78%, ~33x fwd / prior peak).
• Kill (fundamental): Azure growth <20% for 2 qtrs · gross margin <~66% (Stifel flag) · AI capex rising while FCF falls with no monetization (ROIC keeps sliding) · buyback/dividend halt.
• Catalyst: Q4 FY26 earnings Jul 29 (capex/margin print is the crux).

NFLX — $71.56 · core (27.9%) #

• Why: quality compounder (ROE 49%, ROIC 34%, clean BS, op margin 30% and expanding) at a fair (not cheap) price after a -47% drawdown. Trades below the lowest analyst PT ($80). GARP, not deep value. Existing thesis: output/graham/scores/NFLX_thesis.md / NFLX_plan.json.
• Targets: T1 $95 (+33%, ~30x clean fwd — re-assess) · T2 $114 (+59%, analyst mean = thesis re-eval level — trim ⅓) · T3 $135 (+89%, prior 52wk high — trim into strength; analyst high $151). ADD below $70 on further weakness without a thesis break (wider MoS).
• Kill (fundamental): ad revenue misses ~2x 2026 scaling · op margin <25% for 2 non-seasonal qtrs · revenue <high-single-digits FX-neutral 2 qtrs · price hike → material churn · buyback halt + dilutive deal.
• Red flag carried in: 🔴 systematic all-officer insider selling (~$437M, zero buys; shr-002) — mostly 10b5-1, but nobody's buying the dip. 🟡 3 of 4 recent EPS misses.
• Catalyst: Q2 earnings Jul 16 (nearest binary; last quarter's soft guide set a low bar).

PLTR — $107.27 · sliver (19.3%) #

• Why kept (vs INTC cut): unlike INTC there is a floor under the price — fortress balance sheet ($7.8B net cash), positive & accelerating FCF, elite growth (revenue +56–85% YoY, US Commercial +133%), 4 straight beats, and it trades below the analyst mean ($183). You overpay for real hypergrowth, not a hope.
• But: no margin of safety. Fwd P/E 51x, P/S 49x, SBC-adj P/FCF ~93x. So the target ladder IS the risk control — you trim mechanically into strength (no price stop allowed; 52% vol means a stop would whipsaw).
• Targets (trim ⅓ at each): T1 $150 (+40%, ~70x fwd) · T2 $185 (+73%, analyst mean) · T3 $210 (+96%, prior high / ~100x fwd bubble multiple — exit most of the sliver unless fundamentals grew into it).
• Kill (fundamental): US-Commercial growth decelerates sharply (the engine) · SBC re-accelerates as % of FCF · revenue growth <30% while still >40x fwd · guidance cut.
• Red flag carried in: 🔴 all C-suite sold (~$135M, zero buys; shr-002, mostly 10b5-1). 🔴 excluded by the quality factor (§5). Pelosi context: a verified Jun-23 disclosure shows 200 deep-ITM INTC LEAPS (not PLTR) — irrelevant to this name; noted only so a future session doesn't conflate the two (shr-026: lead-gen only).
• Catalyst: Q2 earnings ~Aug 4.

3. Phasing — respect the catalyst wall #

All three report inside ~6 weeks: NFLX Jul 16 · MSFT Jul 29 · PLTR ~Aug 4. Two ways to run it:

• One-shot: deploy §1 now — justified because the weighting already de-risks (MSFT anchor, PLTR sliver) and you're a long-horizon (5y+) holder.
• Two-tranche (preferred): T1 now ≈ MSFT (5) + NFLX (13) ≈ €2,300 (best MoS/quality). Hold PLTR's ~€550 + buffer for T2 after the prints — add PLTR once Q2 resolves the binary, or add to MSFT/NFLX on further weakness without a thesis break (average down at wider MoS; shr-034). PLTR is the most deferrable leg.

4. Sizing method — academic + Graham (the audit trail) #

• No mean-variance optimizer. DeMiguel-Garlappi-Uppal (2009): with ~3 assets and noisy return estimates, optimization loses to naive weighting OOS. Build from risk, tilt by replicated factor + Graham MoS, cap hard.
• Base = inverse-volatility (Maillard-Roncalli-Teïletche 2010; Asness-Frazzini-Pedersen 2012). σ: MSFT 26%, NFLX 34%, PLTR 52% → equal-risk weights ≈ 44/34/22.
• Tilt = replicated factors, all aligned: Quality (QMJ — AFP 2019; gross-profitability — Novy-Marx 2013; Buffett's Alpha — FKP 2018) and Betting-Against-Beta (Frazzini-Pedersen 2014) + low-vol both say overweight MSFT, keep PLTR small. Graham growth-addendum (shr-004): MSFT +32% MoS; NFLX thin/fair; PLTR none → the freed INTC budget went to MSFT, and PLTR stays a sliver.
• Momentum caveat (Jegadeesh-Titman 1993): all three are in negative price momentum (near 52wk lows) — these are contrarian/mean-reversion entries against the momentum factor. Eyes open: you're buying weakness.
• Kelly go/no-go (shr-006, shr-010): MSFT yes (biggest), NFLX yes (moderate), PLTR borderline (small). Fractional-Kelly / parameter-uncertainty (shr-006) keeps any single name from being over-bet on a noisy forecast.

5. Comparison — the 3-stock sleeve vs IS3Q (the charter-compliant quality vehicle) #

IS3Q = iShares Edge MSCI World Quality Factor UCITS ETF (Acc), ISIN IE00BP3QZ601. TER 0.25%, AUM ~€4.8B, 294 holdings, US 73%, IT 31%, beta 0.95 vs MSCI World, 5yr σ ~12.4%. Tracks MSCI World Sector Neutral Quality, scoring each name on three pillars: ROE, low debt/equity, low earnings variability.

The key overlap finding:

Head-to-head:

Honest verdict. IS3Q is the academically-favored, charter-compliant way to own the quality edge — diversified, cheap, lower-drawdown — and it already gives you MSFT (4.3%) + NFLX (1.1%). Note quality has lagged plain MSCI World over 5–10yr (it's a risk-reducer, not a return-maximizer). The 3-stock sleeve is not a quality-factor play — it's a concentrated conviction bet that these names, bought at these drawdowns (MSFT -36%, NFLX -47%, PLTR -49%), beat the diversified factor. Two legs (MSFT, NFLX) are legitimate quality-at-a-discount (GARP); the third (PLTR) is an explicit anti-factor growth bet — which is precisely why it's the smallest sliver. The sleeve only wins if the entry-timing conviction pays off beyond what the factor captures (a value/timing bet, per DeMiguel hard to prove OOS). If you ever want to de-risk: IS3Q is the swap — it keeps ~5% of this exposure, diversified, for 0.25%. The sleeve as built is the deliberate higher-variance choice, sized down accordingly.

6. Same-day red-flag check (shr-020, 2026-06-26) #

• 🔴 Insider selling — NFLX & PLTR: every senior officer sold, zero open-market buys (shr-002; mostly 10b5-1).
• 🟡 Earnings wall: three binary prints in ~6 weeks (Jul 16 / Jul 29 / Aug 4).
• 🟡 Weak internal diversification: three correlated tech names; β≈1.29, σ≈28%. Concentrated, sized down by quality/vol.
• 🟢 Balance sheets: all three financially sound (MSFT/PLTR net cash; NFLX clean, net debt 0.4x).
• 🟢 INTC removed (was the negative-MoS, above-PT name) — sleeve quality improved by the cut.
• Existing related alerts reconciled: MSFT below 350 (triggered, dip-entry), MSFT below 310 (strong value), PLTR below 70 (candidate-on-deck). My TP ladder layers on top (all "above", separate keys) — no conflict.

7. Watchlist alerts set (2026-06-26) #

Monitoring calendar: NFLX earnings Jul 16 · MSFT Jul 29 · PLTR ~Aug 4. Check live price + EUR/USD in DEGIRO before any fill. Targets are USD; positions are USD.

Sources: same-day yfinance + Finnhub agent pulls 2026-06-25/26. IS3Q: MSCI World Sector Neutral Quality factsheet (May 29 2026), iShares/StockAnalysis/JustETF. Pelosi: House Clerk PTR 20033725 (Jun 23 2026). NFLX detail in NFLX_thesis.md / NFLX_plan.json.