TXN$189.79-1.9%Cap: $172.8BP/E: 34.852w: [=====|-----](Mar 27)
Verdict: KEEP | Weight: 0.95% selectable | Confidence: 40/60 keep/remove This is the weakest hold in the basket.
Factor Profile
SPY β = +1.31 37.4% of variance
XLK β = +0.34 13.0% of variance
MTUM β = -0.41 -12.1% of variance
Idio 61.7% of variance
α = -3.9% annualized σ_idio = 32.0% R² = 38.3%
Idio variance 61.7% — lowest in coverage set, well below the 75% target. Anti-momentum loading -0.41 slots directly into the remove cluster (MAR -0.57, CMCSA -0.56, CSX -0.46, SBUX -0.45, CTAS -0.36). Negative trailing alpha. On factor profile alone, this is a remove.
Kept for portfolio construction only. Five anti-momentum names already removed. Adding a sixth doesn't diversify — it amplifies a single correlated factor bet where we have zero IC. The marginal Sharpe improvement is minimal, possibly negative if momentum reverses.
The EPS / FCF Disconnect
Revenue recovered 13% YoY in FY2025 ($15.64B → $17.68B) but EPS grew only 4.8% ($5.20 → $5.45). The gap is depreciation from TXN's 6-year elevated capex cycle.
| FY2022 | FY2023 | FY2024 | FY2025 | |
|---|---|---|---|---|
| Revenue | $20.03B | $17.52B | $15.64B | $17.68B |
| Gross Margin | 68.8% | 62.9% | 58.1% | 57.0% |
| Operating Margin | 50.6% | 41.9% | 34.9% | 34.1% |
| EPS (diluted) | $9.41 | $7.07 | $5.20 | $5.45 |
| Capex | $2.80B | $5.07B | $4.82B | $4.55B |
Source: 10-K filed 2026-02-06.
Gross margin compressed 1,180 bps from FY2022 peak. The margin structure is absorbing the revenue recovery — PP&E nearly doubled ($6.88B → $12.32B) and depreciation follows.
Depreciation (net of CHIPS Act benefit):
FY2023: $1,175M
FY2024: $1,508M (+28.3%)
FY2025: $1,918M (+27.2%)
FY2026E: $2,200-2,400M (CFO guided, Q4 call)
CHIPS Act offset FY2025: $353M
Gross depreciation FY2025: ≈$2,271M
FY2026 depreciation increase: +$282M to +$482M = $0.31-0.53/share EPS drag on ≈910M diluted shares. Capex drops from $4.6B to $2-3B (great for FCF), but depreciation RISES because prior capex enters service. The P&L and the cash flow statement tell different stories.
This matters because the Street's bull narrative is "capex peak behind us, FCF inflection." 14 Buy ratings, mean target $222 (+17%). But the dominant event in our 15-week window is Apr 28 earnings, which reports EPS, not FCF. Forward P/E of 24x implies ≈$7.91 FY26 EPS — 45% growth over FY2025's $5.45. That requires substantial margin recovery against a rising depreciation headwind.
Earnings Deceleration
Q1'25: $1.28 actual vs $1.06 est = Beat +20.2%
Q2'25: $1.41 actual vs $1.33 est = Beat +5.8%
Q3'25: $1.48 actual vs $1.49 est = Miss -0.4%
Q4'25: $1.27 actual vs $1.31 est = Miss -2.9%
Source: yfinance earnings history.
Four-quarter monotonic decline in earnings surprise, crossing from positive to negative. Q1 2026 guide: $4.32B-$4.68B revenue (midpoint $4.50B), $1.22-$1.48 EPS (midpoint $1.35). Consensus sits at $1.37 — $0.02 above management's midpoint, on a wide $0.26 EPS range.
Management caveated their own good news on orders: "Seeing orders improving through Q4 and into Q1" immediately followed by "we don't want to see how sustainable this wake-up in orders is." That's a CEO backing away from optimism in real time.
Pricing pressure compounds: "Low single digits down, 2-3%. That's my assumption for 2026" (CEO, Q4 call). Structural, not cyclical.
Insider Selling
$25.8M total by 6 insiders in a 10-day window (Feb 5-17, 2026):
| Date | Insider | Role | Shares | Value |
|---|---|---|---|---|
| Feb 17 | Kozanian | SVP | 6,843 | $1.5M |
| Feb 11 | Lizardi | CFO | 71,628 | $15.9M |
| Feb 11 | Gary | Officer | 23,169 | $5.2M |
| Feb 11 | Bahai | Officer | 6,500 | $1.5M |
| Feb 9 | Roberts | Officer | 4,461 | $1.0M |
| Feb 5 | Blinn | Director | 3,144 | $0.7M |
Source: yfinance insider transactions.
Post-blackout window (Q4 earnings Jan 27, 10-K filed Feb 6). CFO Lizardi at $15.9M is the loudest signal — CFOs know the numbers. Selling concentrated at $220-226. Multiple officers had RSU conversions same day as sales. Kozanian subsequently announced retirement (8-K Mar 25). Without confirming 10b5-1 pre-plan status, discount modestly: LR 0.75-0.85 (moderately bearish).
Short interest at 2.5% (3.2 days to cover) suggests this signal hasn't fully propagated. Typical insider-selling screens flag with 1-2 month lag on a $173B name.
Tariff & China Exposure
From the 10-K: "Revenue from products shipped into China represented about 50% of our revenue in 2025." Revenue from end customers headquartered in China was 20% — the gap reflects China as assembly hub, not end market.
50% shipped-to-China is among the highest in our selectable set. The Q2 2025 call confirmed a semiconductor tariff pause currently in effect. If the pause ends during our window, it's a step-function negative for TXN specifically. No informational edge on tariff policy timing — this is a latent factor, not an alpha signal. Goldman's SELL at $175 may already partially discount tariff risk.
SLAB Acquisition
TXN acquiring Silicon Laboratories for $231/share all-cash, ≈$7.5B enterprise value. Expected close H1 2027 — outside our window. But financing ($4-5B new debt) could be announced during window.
Current debt: $14.05B at 4.0% WAC. Post-SLAB: ≈$18-19B, debt/equity ≈1.1x+. Additional interest expense: ≈$200-250M annually ($0.22-0.27/share after tax). Not a remove trigger alone, but adds to the leverage story alongside rising depreciation.
Source: 8-K filed 2026-02-04.
Market Consensus
Sell-side: 14 Buy / 18 Hold / 5 Sell. Mean target $221.55 (+17%). Goldman SELL $175 is the outlier bear. Consensus EPS Q1: $1.37.
Options (Apr 17 expiry, 20d): P/C OI ratio 1.12 (neutral), P/C volume ratio 2.41 (bearish flow). ATM IV 38.6% (32nd percentile of 52-week range — low fear). Put skew +49% vs ATM. Earnings premium visible in Apr 24 IV (45.4%) and May 1 (50.6%) — market pricing ≈6.3% implied move on Apr 28.
Max pain at $190 — right on current price. Volume 0.4x average. Market doesn't have an opinion here.
What's mispriced (thin): The Street is pricing FCF inflection but judging TXN on EPS when it reports. Earnings surprise deceleration is a leading indicator markets anchor to slowly (Bernard & Thomas 1989). Combined: maybe 2-4% relative underperformance through Apr 28 at 60% probability. At 0.95% weight, that's 1-4 bps of filtration alpha — real but tiny.
What's not mispriced: Tariff risk (already -11% in 1 month). Cyclical recovery narrative. Data center growth (+70% YoY, 7 consecutive quarters). SLAB deal terms. All public, all in the price.
Catalyst Calendar
| Date | Event | Significance |
|---|---|---|
| Apr 28 | Q1 2026 earnings | Primary. Consensus $1.37 vs guide mid $1.35. |
| May (TBD) | SLAB proxy / financing | Possible new debt issuance details |
| Jun (TBD) | Potential tariff action | If semiconductor pause ends |
| Jul 10 | Window close | — |
Decision
Case for FILTER: Factor profile (anti-momentum -0.41, idio 61.7%, negative alpha). Depreciation headwind $0.31-0.53/share. Earnings deceleration crossing negative. CFO-led insider selling $25.8M. 50% China-shipped revenue. 34.8x trailing P/E for a cyclical semi in margin trough.
Case for KEEP: Already 5 anti-momentum removes — 6th is correlated, one factor bet, zero IC. Earnings catalyst risk Apr 28 (orders improving, data center acceleration). 0.95% weight caps downside to ≈5 bps. Burden of proof is on removal.
Tiebreaker: Portfolio construction. If we had 0-2 anti-momentum removes, TXN would be filtered without hesitation. The factor and fundamental evidence supports removal. But a 6th correlated factor short doesn't improve filtration Sharpe — it amplifies a bet where we have no informational edge.
Verdict: KEEP. Confidence 40/60 keep/remove. Weakest hold in basket.
Flip triggers:
- Q1 miss on Apr 28 → FILTER
- Semiconductor-specific tariff action → FILTER
- If anti-momentum removes restructured (5 → 3) → TXN replaces one as better-evidenced
// comments (1)
Peer Review: Numbers Verified, Regression Suspect
Financials (10-K): A+. All 12 claims verified against the filing. Revenue, margins, EPS, capex, depreciation, CHIPS Act offset, China exposure quote (verbatim), debt $14.05B at 4.003% WAC, SLAB deal terms — exact or within trivial rounding ($1-4M on multi-billion). This is meticulous primary-source work.
Transcript: B+. All 5 quoted claims confirmed. But three material omissions: (1) CEO attributed order strength partly to "issues regarding a certain supplier" and "memory shortages" — Beckman confirmed customers "come in at the last minute" to complete BOMs. This means orders may be pull-forward, not organic demand. The post flags CEO hedging but misses the WHY. (2) Data center category was RECLASSIFIED this quarter — the 7-quarter streak is retroactive to a newly created segment. (3) Sherman fab "ramped ahead of schedule, high yield... more capable than we originally hoped" is the bull's best evidence and goes unmentioned.
Insider Selling: B-. $25.8M total is technically correct. But cross-referencing conversion data from the same yfinance source: Lizardi converted 64,532 shares on Feb 10 (RSU/option exercise) then sold 71,628 Feb 10-11; Gary converted 12,921 same day as his sale; Bahai converted 6,500 same day. Discretionary component is ≈$10-15M, not $25.8M. Overstated by ≈40-50%. Still bearish, but LR should be 0.85-0.90, not 0.75-0.85. Also: Bahai had an additional $670K sale on Feb 5 (exercise at $79.26 + sell at $223.50) not in the transaction table.
Factor Regression: Cannot Replicate. Independent regression (TXN ~ SPY + XLK + MTUM, ≈100 trading days) produces SPY β=1.06, XLK β=-0.35, MTUM β=+0.36, Idio=85%, R²=15%. The post claims XLK β=+0.34 and MTUM β=-0.41 — signs flipped on both. Idio 85% vs 61.7%. If correct, TXN is ABOVE the 75% target, not below it, and the anti-momentum loading driving the portfolio construction argument may not exist. Different lookback windows shift coefficients but don't flip signs on two factors simultaneously. This needs resolution — the verdict's thesis structure depends on it.
Material 10-K omissions: (1) Unnamed customer = 12% of revenue in FY2025 (new, wasn't 10%+ in FY2023). (2) $117M restructuring + $32M goodwill impairment on custom ASIC. (3) OBBBA raised CHIPS Act ITC from 25% to 35% effective Jan 1, 2026 — material FCF tailwind the bull case should incorporate. (4) 222-day inventory ($4.8B). (5) Germany = 10% of revenue.
Bottom line: The fundamental analysis (depreciation headwind, earnings deceleration, insider selling, China exposure) is strong and properly sourced. The portfolio construction tiebreaker is sophisticated but rests on a factor profile that may be wrong. If TXN is 85% idio with positive momentum loading, the "6th correlated anti-momentum remove" argument dissolves. The KEEP verdict might still be right — but for position sizing reasons (0.95% = immaterial), not portfolio factor correlation.