TXN

Texas Instruments

A quality-led, cycle-recovery, cash-generative core long whose main job in this cohort is to neutralize the two structural risks the rest cannot hedge: Taiwan-tail concentration and AI-capex single-factor concentration. The screener's +1 long is intact, but the easy money is gone — the asymmetric trade was at $150–$180 in late 2024. At $280, the equity now trades at 14.4× EV/Sales / 34× forward P/E with the bull case substantially priced in. Buy weakness, not multiple. Add aggressively into any sub-$240 reset that the September 13 2026 MOFCOM final determination is likely to deliver.

This does NOT clear the 2:1 bar that justified medium sizing at $180. At $280 the asymmetry is adequate for a medium position only because (a) the dividend is real cash, (b) the cohort hedge value is structural, and (c) the bear case downside is bounded by quality. The asymmetric trade was 12–18 months ago. Don't kid yourself about that.

TXN's GaN device share is only ~5–7% globally (#5–6 per Yole). The vertical-integration thesis does not require TXN to win discrete GaN device share. It requires TXN to capture more system-BOM dollars by feeding internal GaN into the PMIC + isolated bus converter + multiphase buck stack TI debuted at NVIDIA GTC March 2026 — the broadest-stack 800V offer at 97.6% efficiency, >2,000W/in³.

The cohort-relevant read: if the GaN four-way race goes to scale (Innoscience IDM cost curve dominates), TXN's vertical-integration moat at the system-BOM layer survives because it's cost-structure-based, not patent-based. NVTS's density-mindshare moat erodes faster than TXN's vertical-integration moat. The base case at the April 2 ITC FD (Innoscience prevails) is thesis-supporting for TXN within the GaN four-way race.

FY25 revenue +13% · Q1'26 +19% YoY (industrial +30% · data center +90%) · capex peaking at $4.6B and cut to $2–3B 2026 guide · FCF/share recovering from the $1.47 trough toward the $8+ management 2026 target. Confirmed in disclosed prints, not forecasts.

MOFCOM antidumping investigation on US analog imports. Bernstein sizes ~11.4% revenue exposed; we estimate 3–6% of FY26 revenue at meaningful adverse-outcome risk. Cumulative 51.77% Chinese price decline 2022–2024 cited as injury. Highest-margin commodity-analog slice (Personal Electronics + low-end industrial) is the target.

Catalyst-aware sizing protocol: hold TXN at 1/2-to-2/3 size through Q3 2026. Defend through the binary. Scale back to full only after the September 13 determination lands. Add aggressively into sub-$240 reset that the FD is likely to deliver. The Q4 binary is a defensive-posture quarter — consider a put hedge into the September FD if the implied vol cheapens, or trim 25% in late August 2026.

1. MOFCOM punitive outcome on September 13, 2026. Mechanism: duty rate >20% with retroactive effect on Q4 2024+ invoicing. Chinese OEM procurement pivots to SG Micro / Silergy / 3PEAK / Bright Power / Awinic on accelerated qualification. Structural rather than cyclical step-down — qualification cycles 6–18 months for industrial, 18–36 months for automotive. Action: cut to 1/3 size immediately on adverse FD.
2. 300mm cost-out under-delivers; SM1 utilization stalls. Mechanism: SM1 entered HVM December 2025 at low single-digit utilization. Depreciation loads regardless of utilization. If embedded-analog cycle stalls 2H'26, SM1 utilization stays below 50% through 2027. Promised 350bps of GM tailwind under-delivers (only 100bps materializes). FCF/share stalls at $5–6. Watch: Q3'26/Q4'26 SM1 utilization disclosure below 40% + capex re-creep above $750M/quarter + GM stuck below 59%.
3. Hyperscaler ASIC integration consumes peripheral analog content. Mechanism: hyperscalers integrate the supervisory, sequencing, hot-swap, and isolation analog into custom silicon (NVIDIA absorbing PMIC into BlueField/Vera; AWS Trainium / Google TPU same). TI's per-rack content compresses 30–50% by 2028. Watch: NVIDIA Vera Rubin Ultra (2027 timeline) reference architecture · Meta Catalina / Microsoft HVDC roadmap · any disclosed TI design-loss on a major hyperscaler reference platform.

TXN is the cohort's primary macro hedge AND primary Taiwan-tail hedge. This is the single most important cohort-construction finding from the C-TXN-1 expansion run. State this without burying it: the cohort's biggest unhedged risk before this analyst cycle was Taiwan-tail concentration (synthesis Open Question §1) — and adding TXN materially closes that gap.

The honest cohort frame: TXN provides genuine cohort-hedge value on the four largest cohort-shared risks (Taiwan-tail · AI-capex · rate duration · recession), but it is long China structural — and that exposure concentrates in a single binary date (September 13 2026 MOFCOM). Own TXN at medium size, but don't let the binary date catch the position naked.