Verdict: KEEP — Mega-Cap Anchor, No Active Bet

Mega-cap anchor (#5 in QQQ, 3.81% weight). Held at benchmark weight by construction. Zero informational edge on a $1.36T, 48-analyst name. Q1 earnings April 21 is a known landmine — European demand destruction, FSD transition, Model S/X EOL — but consensus already reflects most of it. Implied earnings move of ±6.1% looks cheap vs. historical, but not our trade.

Factor Decomposition

250-Day Trailing (iev regress)

Alpha = +14.0% annualized. Sigma_idio = 40.0%. Idio Sharpe = 0.35. R² = 0.48, n = 250.

15-Week Window (Dec 16 2025 - Mar 26 2026, n=69)

Four regression models tell a consistent story:

Model 1 — TSLA ~ SPY: beta = 1.85 (t=6.84), alpha = -51.9% ann (p=0.329, not significant), R² = 41.1%, idio var = 58.9%.

Model 2 — TSLA ~ QQQ: beta = 1.43 (t=7.19), alpha = -53.2% ann (p=0.308, not significant), R² = 43.5%, idio var = 56.5%.

Model 4 — TSLA ~ SPY + XLY: beta_SPY = +0.47 (t=1.13, NOT significant), beta_XLY = +1.21 (t=4.08, significant), R² = 53.0%, idio var = 47.0%.

The XLY finding matters: when you add consumer discretionary, SPY beta becomes insignificant. TSLA is trading as a consumer discretionary stock, not a tech/momentum name. XLY was -10.42% in the window vs QQQ -5.90%.

Regime Change

The 250d trailing shows alpha = +14.0% (reflecting the 2025 momentum run from $214 to $499). The 15-week window shows alpha = -53.2% annualized, with -13.29% idiosyncratic underperformance vs QQQ. The regime has changed. Trailing alpha reflects stale momentum; window alpha reflects the current reality.

Window Return Attribution

Comparables: ARKK -14.92%, XLY -10.42%, SPY -4.70%, MTUM -3.73%.

Assessment

52-57% idio variance across all models, well below the 75% threshold. TSLA is half a market bet with a stock ticker. Any active bet on TSLA is substantially a market timing bet. We don't time markets.

European Demand Destruction

Q1 2026 European registration data (aggregated from 20+ public automotive trackers):

The -45% vs +24% divergence is Musk-specific brand damage — DOGE involvement, German political gesture, consumer boycott movement. Not stale lineup or EV competition alone.

What CFO Taneja said on the Q4 call (Jan 28): "Record deliveries in smaller countries... bigger backlog than in recent years." What happened: Norway -88%, Europe overall -45%. Q4 momentum did not carry into Q1.

Tesla is responding with 0% financing and dealer contributions in UK. Giga Berlin reported ≈1,700 layoffs in the period.

This is the key Q1 risk: April 21 earnings will be the FIRST report capturing the full European demand destruction.

Market Consensus — Detail

What the Street Expects

Earnings Track Record

Three of four quarters missed. The one beat (Q4) came with gross margin recovery above 20% for the first time in two years.

Options-Implied Earnings Move: ±6.1%

Derived from term structure isolation:

• Pre-earnings expiry (April 17, 20d): IV = 46.4%
• Post-earnings expiry (April 24, 27d): IV = 50.3%
• Incremental variance: (0.503² x 27/365) - (0.464² x 20/365) = 0.00692
• Strip 4 regular trading days at pre-earnings daily variance: earnings event variance = 0.00377
• Implied earnings move: sqrt(0.00377) = 6.1%

That's cheap for TSLA. IV sits at 23rd percentile of its 52-week range (30.9%-97.7%). The market is not expecting fireworks — which is notable given the headwind stack and 75% miss rate. Stock is already -13.3% in the past month, so the options market may be pricing "bad news already absorbed."

Skew: 3:1 Downside

Unusual put activity: 58 put strikes with vol > 2x open interest on April 17 expiry. Notably, $200 puts traded 10,386 contracts vs. 3,332 OI (3.1x) — institutional crash insurance.

Not a Consensus — A Civil War

The $25-to-$600 target range (24x spread) tells you the Street has not converged on what this company IS:

• Bull camp (23 analysts, $415-$600): Pricing robotaxi + Optimus + AI optionality. Wedbush $600, RBC $500, Cantor $510.
• Neutral camp (17 analysts, $300-$460): Car business + maybe AI. Morgan Stanley $415.
• Bear camp (8 analysts, $25-$125): Car business only, overvalued. Wells Fargo $125, GLJ $25.

The $421 mean target is a mathematical artifact — nobody actually thinks TSLA is worth $421. The real disagreement: is this a car company (10-15x on $1.08 EPS = $10-$16/share on auto alone) or a technology platform (current $362)? The ≈$1.2T gap is optionality on multi-year moonshots.

Mispricing Assessment

Edge = P_you - P_market. On this name, the gap is negligible.

Our 60% Q1 miss probability vs. market-implied ≈50-55% (inferred from: 13% price drawdown, heavy put skew, lowered delivery estimates, oversold RSI). That's ≈5 percentage points of edge. Maybe. It's noise.

Every input to our thesis — European registration data, FSD transition mechanics, S/X EOL, tariff quantification — is public domain, covered by 48 analysts at a $1.36T market cap. We found no evidence the market hasn't already processed.

The only potential second-order mispricing: margin compression from European desperation pricing (0% financing, dealer contributions) hitting both revenue-per-unit and margin-per-unit even if delivery numbers roughly hit consensus. Plus FSD deferred revenue timing and S/X fixed cost deabsorption. Combined, maybe $0.05-0.08 EPS downside not fully captured in the $0.40 estimate. But this is model uncertainty, not informational edge.

Reflexivity check: Stock price DOES affect TSLA fundamentals (cheap equity → capital raises → fund moonshots). Fading the narrative here is fighting a self-reinforcing loop on the bull side. On the bear side, brand damage → demand destruction → lower revenue → lower stock is also reflexive. Both loops are active. Neither is our trade.

Q1 Earnings Setup (April 21)

Eight Headwinds

1. European demand destruction — First full quarter of -45% European decline while EV market +24%. Not visible in Q4 data.
2. FSD subscription transition — First quarter of subscription-only. 70% of ≈1.1M FSD customers were upfront ($8K-$12K). Shifts to monthly recognition.
3. Model S/X end-of-life — Production ending Q1. Mix shift to lower-ASP vehicles. Fixed cost deabsorption at Fremont.
4. Tariff drag — >$400M/Q across both segments (quantified from Q3 transcript), ≈$200M auto alone.
5. Operating expense ramp — Q4 opex +$500M sequentially. "Expect this trend to continue for full year 2026."
6. Energy margin compression — Management pre-guided: "increased low-cost competition, policy uncertainty, tariff costs."
7. CyberCab costs without revenue — Production starts April. Costs already hitting P&L. No meaningful Q1 contribution.
8. $20B+ capex ramp — Six new factories starting. FCF likely negative.

Four Tailwinds

1. Q4 gross margin recovery — >20% first time in 2 years. Can it sustain?
2. New lower-priced models — "Launched least expensive models ever" per Lars Moravy.
3. Energy storage momentum — Backlog strong, MegaPack 3 launching.
4. Robotaxi narrative — Geographic expansion updates, positive sentiment even without revenue.

Prediction

60% probability TSLA misses $0.40 consensus (pred-jypan6). Basis: headwind stack outweighs tailwinds, 75% miss rate over last 4 quarters, European demand destruction first fully visible in this report. The bar is low ($0.40 vs Q1 2025's $0.27) but the headwinds are stacking simultaneously. Resolves April 22.

Financials (FY 2025)

Revenue declined 3% YoY. Auto revenue -10%. Energy storage +27% ($12.8B) is the growth engine. Operating margin compressed 1,220 bps over three years (16.8% to 4.6%). EPS down 75% from FY2022 peak. Balance sheet clean: $16.5B cash, no going concern risk.

Segment mix shifting: energy margins (29.8% Q4 gross) now exceed auto margins (17.9% Q4 ex-credits). If energy continues at +25-30% growth, it becomes the margin story.

Valuation

Trailing P/E 335x, forward P/E 129x. Auto business at peer multiples (15-20x on ≈$3B auto operating income) = $45-60B. Energy storage at growth multiples (30-40x on ≈$3.8B gross profit) = $100-150B. Total "tangible" value ≈$150-200B. The remaining ≈$1.2T is optionality on robotaxi, Optimus, AI, TerraFab, solar. Outside our 15-week window.

Insider Activity

No insider buying. $64M combined director-level selling over 3 months. Directionally negative but not unusual for a $1.36T name with elevated stock-based comp.

Technicals

Below both 50-day MA ($415) and 200-day MA ($394). RSI 32.8 (oversold). 52% of 52-week range ($214-$499). Unusual put activity (58 strikes with vol > 2x OI) despite bullish P/C ratio (0.52). Max pain at $400 (10.5% above current).

Basket Impact

At 3.81% weight with beta_QQQ = 1.43, TSLA amplifies QQQ moves by 43% at its weight. In the 15-week window, TSLA dragged QQQ by ≈83 bps beyond its benchmark (-21.71% x 3.81%).

TSLA is trading as consumer discretionary (beta_XLY = 1.21 significant, SPY insignificant when both included). QQQ's TSLA exposure is really consumer discretionary exposure in a tech index. If consumer discretionary weakens further (tariffs, confidence), TSLA drags QQQ disproportionately. If tech rallies but consumer doesn't follow, TSLA lags the index.

57% idio variance in window means a TSLA-specific crisis (CyberCab delay, Musk escalation) affects ≈2.2% of portfolio (3.81% x 57%) as pure idio risk — material but not fatal.

Window Catalysts (March 27 - July 10)

Why Not FILTER

Could you argue for removing TSLA? The case: -13.29% idio residual in window, European demand -45%, 60% miss probability, regime change from +14% trailing alpha to -53% window alpha, insider selling, RSI oversold with no buying support.

The rebuttal is definitive:

1. 3.81% weight. Even at our 60% miss probability, expected filtration alpha from removing TSLA = 3.81% x (expected TSLA underperformance vs QQQ). At ±6.1% implied move, this is ±23 bps — within noise for the basket.
2. Zero edge. 48 analysts. $1.36T market cap. Everything we know, the market knows. Edge = P_you - P_market ≈ 5 pts. Not tradeable.
3. 52% idio variance. Half the variance is market beta. Shorting TSLA is half a market short. We'd need to know what the market does, not just what TSLA does.
4. Reflexivity cuts both ways. CyberCab production starts April. Robotaxi expansion ongoing. Optimus 3 unveiling. Any of these could flip the narrative in a name that runs on stories, not earnings. The right tail is real.
5. Mega-cap anchor by construction. Top 10 by QQQ weight. The filtration framework holds mega-caps at benchmark weight because the information environment is too efficient to generate edge.

Read-Throughs for Active Positions

• AI capex: $20B+ TSLA capex validates AI infrastructure demand (AVGO thesis)
• Energy storage: Margin compression guidance is negative for utility valuations
• Consumer discretionary: TSLA trading as XLY proxy — monitor for basket factor exposure
• European EV: Brand damage creating opportunity for competitors (not in our basket)