Texas Instruments Incorporated (NASDAQ: TXN) — The 300mm Bet Is Built; the Price Already Assumes It Pays Off

Independent research note. Prepared 2026-06-10. As-of price ~$282.

This note discusses valuation only as embedded expectations and scenarios. The analytical body contains no buy/sell recommendation and no price target. The single, deliberate exception is the Claude’s Take block immediately below, which is fenced off as the author’s subjective view.

⚡ Claude’s Take

This block is the author’s own subjective opinion and general information, not investment advice. Everything below it is position-free analysis.

Verdict: HOLD / great business, full price — accumulate on weakness, do not chase, do not short. Constructive entry zone ~$200–230 (≈ 4% normalized FCF yield); fair ~$250–290; demanding above $300. Conviction: medium.

Texas Instruments is one of the highest-quality industrial-technology franchises in the world — the lowest-cost producer in the best-structured corner of semiconductors, run for two decades around a single honest metric (free cash flow per share), with clean GAAP accounting, trivial stock comp, and a 22-year dividend-growth record. None of that is in dispute, and it is why this is not a short. The problem is entirely price. At ~$282 the stock sits in the richest ~2% of its own ten-year valuation history on price-to-sales (which strips out the cyclicality), trades roughly in line with Analog Devices, and embeds a near-complete recovery: ~$8–9B of normalized free cash flow at a sub-3% forward FCF yield, achieved while depreciation from the new fabs is still rising into 2027 and gross margin sits ~12 points below its 2022 peak. My base case — a successful cyclical and margin recovery toward ~$22B revenue and ~60% gross margin — lands roughly at today’s price. That is the tell: you are being asked to pay today for a payoff the company has not yet delivered, with the bull case requiring revenue (~$25B+) TI has never reached and the bear case (a permanently lower ~57–60% owned-fab margin base, plus China share loss) implying ~−40% with no business failure required. The risk/reward is skewed against the buyer at this price.

The framing is quality-compounder-at-the-wrong-price, not value and not falling knife. What the market has right: capex is genuinely rolling off ($5B → $2–3B), the 300mm cost advantage is real and durable, and FCF is inflecting. What it is pricing generously: the magnitude and certainty of margin normalization, and the durability of the +90% data-center growth and the ~50%-of-revenue China base into an active anti-dumping probe. Single bullish trigger: H2-2026 sustains double-digit YoY growth with the flagged price increases landing and gross margin marching back toward the low-60s% — confirming the FCF/share compounding case and the inflection is real, not a second “false start.” Single bearish trigger: a 2025-style H2 deceleration or a MOFCOM duty/design-out event in China — either would expose that “normalized” FCF is lower and later than the price assumes. Tag: the fabs are built; now the bill — depreciation — comes due before the cash does.

1. Executive Summary

Texas Instruments is the world’s largest analog semiconductor company and one of the last vertically-integrated device makers of scale — it both designs and fabricates the great majority of its ~80,000 products in-house, selling to ~100,000 customers across Industrial (33% of revenue), Automotive (33%), Personal Electronics (21%), Data Center (9%), and Communications (3%). Analog is ~79% of revenue at a ~38.6% segment operating margin and is the entire economic engine; Embedded Processing (15%) has structurally weakened (operating profit fell from ~$1.0B in FY23 to ~$0.30B in FY25). The business model is transactional in form but annuity-like in economics: analog parts ship for 10–20+ years with low obsolescence, automotive/industrial design-ins carry multi-year qualification, and embedded customers reuse their own firmware — so once designed in, a TI part is a multi-year cash stream. Management runs the company explicitly to maximize long-term free cash flow per share.

The investment debate is not about business quality, which is excellent, but about where TI sits in two intertwined cycles. First, a self-imposed capital-spending supercycle: TI spent ~$20B over 2021–2026 building internal 300mm wafer capacity (Sherman TX, Lehi UT, Richardson TX), deliberately ahead of demand, accepting years of under-utilization to secure the industry’s lowest-cost, geopolitically-dependable manufacturing base for the 2030s. That bet crushed free cash flow — from $6.3B (FY21) to a $1.35B trough (FY23) — and FCF per share today (~$2.85 FY25) still sits below its FY21 level, four years on. Capex now rolls off (2026 guide $2–3B vs. ~$5B peak), and FCF is inflecting ($4.4B trailing, vs. $1.7B a year earlier; management calls ~$8/share for 2026 “highly probable”). Second, the analog demand cycle: revenue fell 22% from a $20.0B FY22 peak to a $15.6B FY24 trough, recovered to $17.7B in FY25, and is now growing strongly (Q1-2026 +19% YoY) — but off a deep base, and management itself twice warned that 2025’s strong first half decelerated into a “false start” second half, a pattern that could repeat.

Margins tell the strain: gross margin compressed from 68.8% (FY22 peak) to 57.0% (FY25) and operating margin from 50.6% to 34.1%, driven by new-fab depreciation (rising from $1.18B in FY23 toward $2.2–2.4B guided for 2026, still climbing into 2027) layered on cyclical under-utilization. The capital-return machine kept running through the trough — 22 straight years of dividend increases, dividends reaching $5.0B (FY25) — but at a cost: dividends exceeded free cash flow for three straight years, funded by ~$6B of incremental, trough-priced debt (total debt now ~$14B at a 4% weighted coupon), while buybacks were slashed during the very years the stock derated. In February 2026 TI announced its first sizeable acquisition since 2011 — Silicon Labs, ~$231/share, ~$7.5B enterprise value, all-cash — to shore up the weak Embedded segment with wireless connectivity, adding ~$7B of debt and integration risk for a richly-priced fill-in.

The moat is real and financially proven — economies of scale fused with a genuine 300mm cost advantage (~40% lower die cost than 200mm) and design-in customer captivity, producing a 34% operating margin at the trough and ~30%+ ROE. But it is no longer uncontested: Chinese domestic analog makers (SG Micro, 3Peak, Novosense) are taking structural share in the fragmented low/mid-end, an active September-2025 MOFCOM anti-dumping probe threatens the ~50% of revenue that ships into China, and the owned-fab model may have reset gross margin to a permanently lower base than the capacity-light 2022 peak. At ~$282 — the richest ~2% of TI’s own decade on price-to-sales, ~35x management’s own $8 FCF/share target — the market is pricing the capex roll-off and cost advantage correctly while underwriting a near-complete, durable margin recovery generously. The base case validates the price; the disconfirming risk is the normalized-margin debate and China.

2. Business Overview

What TI is. Texas Instruments designs and manufactures semiconductors — chips that condition, amplify, convert, sense, distribute power, and process data — which it sells to roughly 100,000 customers across more than 30 countries. Founded in 1930 and headquartered in Dallas, TI is the world’s largest analog-chip maker and one of the very few remaining vertically-integrated device manufacturers (IDMs) of scale: it both designs and fabricates the great majority of its silicon in-house, the inverse of the fabless model (Nvidia, Qualcomm, Broadcom) that now dominates the broader semiconductor industry. The portfolio spans more than 80,000 distinct products — an extreme long-tail catalog that is itself a competitive asset (Fact — FY2025 10-K, filed 2026-02-06).

How it makes money — the catalog/long-tail analog model. TI sells physical components, not licenses or subscriptions, so revenue is transactional in form. But the economics behave like recurring revenue because of two mechanics. First, analog parts have unusually long commercial lives — a given part can ship for 10–20+ years with low obsolescence risk, so once a TI device is designed into a customer’s board it generates a multi-year annuity. Second, in Embedded Processing the customer typically writes its own firmware to run on TI’s microcontroller, and many customers prefer to reuse software from one product generation to the next, lengthening the relationship and raising switching costs (Fact — 10-K). The result is a business with no contractual recurring revenue but a revenue base that is sticky, diversified, and resistant to single-customer or single-product shocks: about half of revenue comes from customers outside the largest 50, and management explicitly runs the company to maximize free cash flow per share over the long term rather than near-term growth (Fact — 10-K).

Segments and revenue mix (FY2025, $17.68B total).

TI is, overwhelmingly, an analog company: Analog is ~79% of revenue and the engine of profitability, with Embedded a distinct second business and “Other” a long-tail of legacy/specialty lines. Notably, in FY2025 TI took a non-cash goodwill impairment on its custom-ASIC products inside Other — a small but telling signal that the custom-silicon adjacency has underperformed (Fact — 10-K). This is not an AI-accelerator company; its data-center exposure is power and signal-chain content around AI servers, not the compute itself.

End markets (FY2025, realigned). TI re-cut its market taxonomy in FY2025 to surface its growth priorities:

Industrial and Automotive together are two-thirds of revenue — the two markets with the longest design-in cycles, the highest content-growth trajectory (electrification, factory automation, ADAS), and the stickiest relationships. Management places explicit additional strategic emphasis on Industrial, Automotive, and Data Center as the best long-term growth pools (Fact — 10-K). Data Center, though only 9%, is the fastest-growing line (~+90% YoY in early 2026) and the swing factor in the bull narrative.

The channel shift — direct distribution. A structural change over recent years is TI’s migration to direct sales: more than 80% of FY2025 revenue was direct, including TI.com, with traditional distribution shrinking to a minority role (Fact — 10-K). Direct relationships yield richer demand signal, access to more of each customer’s design projects, the chance to cross-sell more parts into each board, and capture of distributor margin. The trade-off is that TI now carries more inventory and fulfillment burden directly — part of why it builds “ahead of demand” on broad, low-obsolescence parts.

Geography and customer base. Revenue by customer headquarters is US 38%, China 21%, EMEA 21%, Rest of Asia 11%, Japan 7%. Critically, ~50% of revenue is from products shipped into China even though only ~20% is from China-headquartered customers — a function of China’s role as the world’s electronics-assembly hub (Fact — 10-K). No single customer is disclosed as a >10% concentration; the base is genuinely diversified across the 100,000-account, 80,000-SKU long tail.

Verdict. TI is a high-quality, diversified IDM analog-and-embedded franchise whose transactional product sales behave like an annuity because of decade-long part lifecycles, embedded-software lock-in, and an unmatched 80,000-product catalog served increasingly direct. The model is structurally attractive and built deliberately around free cash flow per share — but it is concentrated in cyclical Industrial/Automotive end markets and carries heavy (~50% shipped-in) China exposure that is now a live competitive and geopolitical fault line.

3. Industry Dynamics

Structure: a consolidated-at-the-top, fragmented-at-the-tail oligopoly. The analog and embedded-processing market sits in the structurally most attractive corner of the semiconductor complex — far from the boom-bust commoditization of memory and the winner-take-all capital intensity of leading-edge logic. The analog TAM is roughly $93.7B (2025), and TI is #1 at ~19% share; the top five — TI, Analog Devices, STMicroelectronics, Infineon, NXP — together hold roughly half the market (Fact — Fortune Business Insights / Mordor aggregators, 2025; directional only). Yet TI’s own 10-K is candid that despite consolidation, the analog and embedded markets remain highly fragmented, with competition from dozens of large and small companies including emerging companies, particularly in Asia (Fact — 10-K). This is the key structural nuance: the high-value, high-performance analog sockets are oligopolistic and defensible, while the commodity/mid-low-end tail is fragmented and contestable — precisely where the Chinese entrants are attacking (see the Competitive Position section).

Why barriers are real but not absolute. Unlike leading-edge logic, where the barrier is access to TSMC’s $20–30B fabs and ASML’s EUV monopoly, the analog barrier is a different stack: (a) deep, accumulated process- and packaging-know-how in analog/mixed-signal devices, where performance is defined by design artistry and process maturity rather than the digital node; (b) multi-year qualification cycles in automotive (AEC-Q100, ISO 26262) and industrial that make designed-in parts sticky; and © the catalog breadth and manufacturing scale to serve 100,000 customers economically. These favor incumbents and make share gains slow and expensive — but they are lower barriers than EUV logic, which is why a well-capitalized, state-supported entrant can chip away at the low end.

Profit pools and the cost-advantage dynamic. The defining economic feature of this industry is that profitability concentrates in the lowest-cost, highest-scale manufacturer of broadly-used parts. TI’s central structural lever is 300mm wafer manufacturing: an unpackaged die built on a 300mm wafer costs ~40% less than the same die on a 200mm wafer (Fact — 10-K). Most analog peers still run substantial 200mm/150mm capacity or outsource to foundries; TI is building the industry’s largest internal 300mm base (Richardson RFAB2; the Sherman SM1/SM2 mega-site — up to four fabs, each capable of ~$5–10B of revenue; and Lehi LFAB1/LFAB2 in Utah). This is a genuine, sustained cost advantage that lets TI earn a 34% operating margin at the trough of the margin cycle and, if it chooses, price below sub-scale rivals while still earning a return. It is the analog equivalent of a low-cost commodity producer’s position on the cost curve.

The capital cycle (Marathon lens) — a textbook supply overshoot, mid-correction. Applying Marathon’s supply-side discipline is the single most useful timing frame. The 2021–2023 chip shortage drove the entire group to announce capacity simultaneously: TI’s own $20B+, six-year capex cycle (2021–2026), plus Infineon’s Dresden/Kulim builds, ST/onsemi SiC expansions, NXP/Microchip additions, and a flood of Chinese trailing-edge capacity. That lumpy supply is now arriving into cyclically soft, post-COVID-destocked demand — visible in industry-wide margin compression (TI gross margin 68.8%→57.0%; Infineon segment margin ~27%→17.5% per Infineon’s FY25 reporting). This is the classic capital-cycle sequence: high returns attracted capital, capacity overshot, returns fell. TI is building into the late stage of the cycle, which carries real near-term risk (capex digestion; under-utilized new fabs depressing gross margin via depreciation, not pricing). The mitigants: TI’s incremental capacity is the lowest-cost capacity in the industry (300mm), is partly CHIPS-Act-subsidized (a 35% investment tax credit), and is intended counter-cyclically — TI’s stated bet is to support customers through the cycle and gain share when capacity-constrained rivals cannot supply (Bernstein Strategic Decisions Conference transcript, 2026-05-28, CEO Haviv Ilan). In Marathon terms, TI is positioning to be the survivor that emerges from the correction with the best cost position — but the investor must underwrite a digestion period first.

Cyclicality. The flip side of the secular content-growth tailwind is that analog/industrial/automotive semis are deeply cyclical, driven by inventory swings at industrial OEMs, automotive Tier-1s, and distributors. TI’s 10-K describes the semiconductor cycle of tight supply followed by surplus inventory as an inherent feature, amplified by the long lead time and cost of building fabs (Fact — 10-K). The 2023–2024 downturn took revenue from a $20.0B FY2022 peak to a $15.6B FY2024 trough before a partial $17.7B FY2025 recovery. Investors must underwrite mid-cycle economics, not the peak margins of 2022.

Regulation and geopolitics — now a first-order factor. Three forces matter. (1) CHIPS Act / industrial policy: US subsidies and the enhanced investment tax credit materially lower the after-tax cost of TI’s domestic fab build-out — a tailwind, and the foundation of TI’s “geopolitically dependable” marketing. (2) China export controls and de-risking: bidirectional risk, since ~50% of TI revenue ships into China. (3) The China anti-dumping probe (Sept 2025): China’s MOFCOM opened an anti-dumping investigation into US analog ICs (TI and others), alleging margins of 300%+; analysts warn it could lead to tariffs or usage restrictions, accelerating Chinese customers’ adoption of domestic alternatives (Fact — TrendForce 2025-09-15; SCMP). This is the sharpest structural threat in the industry today and is specific to the high-margin US analog leaders.

Verdict: structurally good industry, with a deteriorating edge. The high-performance analog/embedded core is one of the best industries in technology hardware — consolidated, high-barrier, secularly-growing (electrification, automation, AI-power content), capable of sustained 30%+ operating margins for the cost leader. Three qualifications temper the verdict: the industry is cyclically mid-correction (supply overshoot still digesting); the low/mid-end tail is fragmenting under Chinese entry; and geopolitics has become a structural, not peripheral, risk given the ~50%-shipped-into-China exposure and an active anti-dumping probe. It remains an excellent industry to own the cost-advantaged leader in across a full cycle — but it is no longer the placid, uncontested compounding pond it was a decade ago.

4. Competitive Position

The moat, named (Greenwald). TI is the cleanest example in analog of Greenwald’s strongest configuration — economies of scale fused with customer captivity, reinforced by a structural cost advantage. Each pillar is financially visible and mutually reinforcing:

• Economies of scale. TI is the #1 analog maker globally at ~19% share — the largest volume base in the industry — across which it amortizes fab and R&D fixed costs and an 80,000-product catalog. Scale lets TI run the industry’s largest internal 300mm network that sub-scale rivals cannot economically replicate, and lets it serve the 100,000-customer long tail profitably — breadth no niche competitor can match.
• Structural cost advantage. The 300mm-vs-200mm ~40% lower die cost, layered on largely-depreciated legacy capacity plus new low-cost capacity, gives TI the lowest cost-per-function in volume analog. This is the pillar that distinguishes TI even from high-quality peers: it earned a 34.1% operating margin in FY2025 at the trough of its margin cycle (Fact — 10-K).
• Customer captivity. Analog parts ship for 10–20+ years once designed in; automotive/industrial qualification is multi-year; embedded customers reuse their own firmware across product generations. Re-qualifying a competitor mid-lifecycle means redesign, re-validation, and risk — so customers rarely do it. The market evidence is TI’s durable, decades-long #1 analog position (passes Greenwald’s share-stability test) and through-cycle returns (passes the ROIC test: 34% operating margin and ~30%+ ROE even at trough).

Direct competitive comparison. The peer set splits into the two highest-quality franchises (TI, ADI) and a band of more cyclical, lower-margin IDMs:

The sharpest read: TI and ADI are the two highest-quality analog businesses in the world, and they barely compete head-on — ADI optimizes for the highest-precision, performance-trumps-price sockets (hence its ~65%+ gross margin), while TI optimizes for the broadest, lowest-cost, highest-volume catalog (57% gross margin but far larger scale and the structural cost lever). The remaining IDMs (Infineon, NXP, ST, onsemi, Renesas) are more exposed to cyclical power/automotive and earn structurally lower margins. Against the fabless world, TI’s IDM model is a deliberate cost-and-control choice that pays off specifically in analog, where the device is defined by process/packaging rather than the bleeding-edge node.

Switching costs — real, but graded. Captivity is strongest in automotive and industrial (multi-year qualification, functional-safety certification, decade-long platform lifecycles) and in embedded (firmware reuse). It is weakest in personal electronics and in commodity mid/low-end signal-chain and power parts, where the part is more interchangeable and the buyer price-sensitive. This gradient matters enormously for the China threat: the captive base is well-defended; the non-captive commodity tail is the soft underbelly.

The China threat — pressure-testing the moat. This is the central bear argument and deserves direct treatment. Chinese domestic analog players — SG Micro, 3Peak, Novosense — grew at 2018–2023 revenue CAGRs of ~36%, ~57%, and ~101% respectively (off low bases), targeting mid- and low-end industrial signal-chain parts that overlap directly with TI’s part numbers, with lower prices, local proximity, and a powerful domestic-preference/localization tailwind from US-China sanctions (Fact — SCMP / Findchips, 2025). Two 2025 developments sharpened the threat: (1) China’s Sept-2025 anti-dumping probe into US analog ICs, which could bring tariffs or usage restrictions and explicitly aims to create breathing room for Chinese suppliers; and (2) TI’s own mid-2025 broad price hike on 3,300+ parts, which hands Chinese vendors a pricing umbrella to take share precisely in the contestable tail (Fact — TrendForce 2025-09-15; DigiTimes 2025). With ~50% of TI revenue shipped into China, this is not peripheral. The honest read: a meaningful slice of TI’s low/mid-end industrial revenue shipped into China is structurally — not cyclically — exposed to domestic substitution, and the moat is genuinely eroding at the low end even as it holds at the high end.

TI’s defense, and why the moat is still durable. Three counterweights keep this an erosion at the margin rather than a thesis-breaker. First, the 300mm cost advantage is TI’s weapon back: it can out-price Chinese entrants on broad parts if it chooses, sacrificing some margin to defend share — the mid/low-end is exactly where its cost lever bites. Second, ~80% of revenue is outside China-headquartered customers and the captive auto/industrial/embedded design-in base (with qualification and software lock-in) is far harder to displace than commodity catalog parts. Third, TI’s internalization and US-fab footprint turn the geopolitical split into a marketing advantage for non-China customers seeking dependable supply — and let TI gain share from capacity-constrained Western rivals as the cycle turns.

Where the moat is thin. Two honest caveats. (1) Embedded Processing is a distinct, arguably sub-scale business at $2.7B (15% of revenue) — it competes against larger MCU specialists (Renesas, Microchip, ST, NXP, Infineon), its scale advantage far less commanding than in core analog; the FY2025 custom-ASIC goodwill impairment underscores that not every embedded/custom adjacency has worked, and the Silicon Labs acquisition is partly an admission that organic Embedded execution fell short. (2) Data-center AI-power (9%, fast-growing) is merchant, multi-sourced turf shared with Monolithic Power, ADI, Infineon and Vicor — strong content opportunity, but not the multi-year-qualified captivity of an auto platform; design wins can shift.

Verdict: a durable, multi-pillar moat — but eroding at the low end, not uncontested. TI’s competitive advantage is real, financially proven, and the strongest in volume analog: economies of scale + a genuine 300mm cost advantage + design-in captivity, producing 34% operating margins at a cyclical trough and decades of stable #1 share. The cost-advantaged high end and the captive auto/industrial/embedded base are well-defended. But the moat is no longer un-eroding: Chinese domestic players are taking structural share in the fragmented mid/low-end tail, an active anti-dumping probe and TI’s own price hikes are accelerating that substitution, and ~50% shipped-into-China exposure makes it material. The verdict is durable but contested — a wide moat with a clearly identified, slow leak at the low end that TI has the cost weapon to defend but cannot fully seal.

5. Growth History and Forward Opportunities

The arc: a $20.0B peak (FY22), a $15.6B trough (FY24), and a recovery now running hot off a low base. TI’s revenue collapsed roughly 22% peak-to-trough as the post-COVID analog inventory correction — described by management as “historic” in size and unusually “asynchronous” across end markets (2026-02-24 Capital Management call) — worked through the channel. FY25 recovered to $17.7B and the trailing figure is ~$18.4B, still ~8% below the 2022 peak. The recovery is real but is not a clean secular acceleration: Q1-2026 revenue of $4.8B (+19% YoY, +9% sequential) marked the eighth consecutive quarter of sequential growth, but measured against a deeply depressed FY24 base. Distinguishing the cyclical snap-back from durable content growth is the central task here.

Segment and end-market composition. The more useful cut is by end market, which TI resegmented in early FY2025 to break out Data Center: Industrial ~33%, Automotive ~33%, Personal Electronics ~21%, Data Center ~9%, Communications ~3%. Industrial + Automotive + Data Center are ~75% of revenue, up from 43% in 2013 (2026-02-24 call) — a deliberate decade-long shift toward content-rich, long-lived sockets and away from the legacy mobile/baseband business TI exited years ago.

Organic vs. acquired. TI’s growth has been almost entirely organic for fifteen years — Silicon Labs (announced 2026-02-04) is its first significant acquisition since National Semiconductor in 2011. The profile is therefore not flattered by serial M&A; it reflects share gains, content growth, and capacity-enabled supply reliability. That is a quality positive and a clean read on the underlying business.

Forward drivers — separating secular from cyclical:

• Industrial (cyclical recovery + secular content). Q1-2026 industrial grew +30% YoY and +20% sequentially, “broad across all sectors and regions” — management called it “the tail starting to wake up after a long hibernation” (2026-04-22 call). Crucially, by management’s own math industrial revenue is still ~15% below its 2022 peak, with most sectors (ex-aerospace/defense) running 30–35% below peak. Interpretation: the bulk of the +30% is cyclical replenishment toward a prior peak, layered on a genuine secular content tailwind (automation, sensing, energy efficiency). High-quality, but partly a one-time recovery that will not repeat at this rate.

• Automotive (secular content, cyclically near-peak). Q1-2026 auto grew only mid-single-digits YoY and was flat sequentially, with China down and rest-of-world up. Auto never had the steep correction the other markets did — it has been “steady at an elevated level,” ~1–2 points below its own peak. The secular story (more silicon content per vehicle across BEV/hybrid/ICE — electrification, ADAS, higher-functional-safety nodes) is intact and credible, but because auto sits near peak with no correction absorbed, it carries forward air-pocket risk rather than recovery upside.

• Data Center (the genuinely secular, uncapped engine). The standout: +90% YoY and +25% sequential in Q1-2026, exiting at roughly a $450M+ quarterly run-rate (~10% of revenue). TI sells the analog power tree inside the rack — DC-DC regulators, hot-swap and multiphase controllers/power stages, current sensors, clocks, interface — plus an emerging GaN-based application-specific position for the 800V-DC architecture transition. Is it durable? The demand driver (AI data-center capex and power-density economics) is secular, and TI argues its combination of general-purpose breadth (“tens of thousands of SKUs per rack”), supply reliability when rivals are short, and US/dependable manufacturing is hard to replicate. But: it is only ~10% of revenue, the +90% rate is unsustainable and partly reflects filling competitors’ supply gaps, and power is fiercely contested (Monolithic Power, Infineon, ADI, Vicor). The level is durable; the growth rate is not.

• Silicon Labs / wireless connectivity (acquired, secular). SLAB adds ~1,200 products, a ~15% revenue CAGR since 2014, and ~85% industrial exposure in an early-innings wireless-connectivity space (2026-02-24 call). It is a content-adjacency and cross-sell vehicle — but it does not close until 1H-2027 and contributes nothing to near-term organic growth.

Pricing — a newly favorable variable. After years of low-single-digit annual price declines, pricing was flat YoY and sequentially in Q1-2026, and management flagged price increases as “likely” in 2H-2026 if demand sustains (2026-04-22 call). Flat-to-up pricing on a recovering volume base is a meaningful margin/growth tailwind if it holds.

Verdict: high-quality growth, but a meaningful slice of the current rate is cyclical and non-repeating. The portfolio is positioned in structurally attractive, content-growing end markets (75% industrial/auto/data-center); growth is overwhelmingly organic; pricing has inflected favorably. That is genuinely high-quality. But the headline +19% reported and +30% industrial growth flatter the secular signal — they are measured off a deep trough, industrial is mid-recovery, and data center’s +90% is a small-base rate that will normalize. Management itself raised, unprompted and twice, the risk that 2026 repeats the 2025 “false start.” The honest read: a structurally good business in a genuine cyclical upswing, where the secular content story is real but currently amplified by recovery math.

6. Financial Quality

Verdict up front: Through the cycle, TI is a high-quality franchise — Analog throws off 38–39% operating margins, GAAP earnings are unusually clean, and the balance sheet is investment-grade. But the company is currently being run through the trough of a deliberate, self-inflicted capital-spending supercycle, and the headline numbers reflect that strain. Reported margins compressed materially, free cash flow roughly quartered from its 2021 peak, and FCF per share today sits below its 2021 level. Economics absolutely improve with scale — Analog proves it — but the shareholder has not yet been paid for six years of fab-building. This is a high-quality business in the least flattering phase of its own investment cycle.

Multi-year financial summary (reconciled to filings)

Sources: FY2025 10-K (filed 2026-02-06); prior 10-Ks; Q1-2026 10-Q. FCF/share uses diluted share count. Some FY21–22 debt/depreciation figures are approximate.

The margin story: a real structural step-down, not just cyclical

Gross margin fell from a 68.8% peak (FY22) to 57.0% (FY25) — ~1,200 basis points. Two forces are at work, and separating them is the analytical crux. The first is cyclical under-utilization: the 2023–24 downturn cut revenue ~22% from peak, and a fab-heavy manufacturer with high fixed costs sees gross margin fall faster than revenue. The second, more durable, is the depreciation wave from new 300mm capacity: depreciation rose from $1.18B (FY23) to $1.92B (FY25) and is guided to $2.2–2.4B in 2026, still climbing into 2027. New fabs are placed in service ahead of the demand to fill them, so their depreciation hits the cost line before the revenue arrives. Management attributes FY25 gross-profit pressure to “higher manufacturing costs associated with our planned capacity expansions” (Fact — FY2025 10-K MD&A). Even with revenue recovering 13% in FY25, gross margin still slipped (58.1% → 57.0%) — depreciation overwhelmed the volume recovery. The Q1-26 operating-margin recovery to 37.5% on 18.6% revenue growth shows the operating leverage is intact once factories load; but the 57% gross-margin level is a new, lower base than the 65%+ of last cycle, and a chunk is permanent (more owned-fab depreciation in the cost stack).

Segment reality: Analog is the company; Embedded has broken

The consolidated numbers mask a sharp divergence. Analog generated $14,006M of revenue at a 38.6% operating margin (op profit ~$5,412M) in FY25 — the crown jewel and source of essentially all economic value. Embedded Processing, by contrast, has collapsed: revenue fell from $3,368M (FY23) to $2,697M (FY25), and operating profit cratered from $1,008M (FY23) to $304M (FY25) — an 11.3% margin versus Analog’s 38.6% (Fact — FY2025 10-K segment footnote). Embedded bears a disproportionate share of new-fab depreciation and has lost ground in microcontrollers. Interpretation: the “TI moat” is really an Analog moat; the consolidated 34% operating margin understates how good Analog is and how mediocre Embedded has become — and frames the Silicon Labs deal as a fix for a real problem.

The capex–depreciation–FCF inflection — the whole story in one arc

This is the single most important dynamic in the financials. From FY18–FY22 TI spent ~$0.6–2.8B/yr on capex against ~$8B+ of CFO, minting ~$6B of FCF a year. Then it launched a six-year elevated-capex program to internalize 300mm wafer manufacturing: capex jumped to $5.07B (FY23), $4.82B (FY24), $4.55B (FY25) — far above the long-run ~1.2x-revenue-growth guide. The result: FCF cratered to $1.35B (FY23) and $1.49B (FY24), recovering only to $2.60B (FY25). FCF/share fell from $6.84 (FY21) to $1.47 (FY23), and at $2.85 (FY25) is still below the FY21 level — four years of stagnation in the one metric TI says it exists to maximize.

The bull case is that the inflection is now arriving: management guides 2026 capex to just $2–3B (Q1-26 capex was already down to $676M vs. $1,123M a year earlier), explicitly calling this the end of the elevated cycle. If CFO normalizes to ~$8B mid-cycle against $2.5–3B capex, FCF could re-rate toward $5B+ near-term and higher as volumes build. The skeptical read: (1) that re-rating is prospective, not realized — the shareholder funded six years of building and has yet to be paid; (2) depreciation keeps rising into 2027 even as capex falls, so reported earnings face a continued D&A headwind; and (3) the new fabs only generate the promised FCF if TI can fill them — the bet requires demand not yet in hand. The capex supercycle is a leveraged bet on volume the company does not yet have.

Quality of earnings: genuinely clean, but FCF is flattered by subsidy

TXN’s earnings quality is among the best in large-cap tech. The company reports pure GAAP — no parade of “adjusted” addbacks — and SBC is strikingly low at $419M (FY25), ~2.4% of revenue (versus high-single-digit-percent at most peers), so diluted EPS is not quietly inflated by stock-comp exclusions. Net income ($5.00B) tracks CFO ($7.15B) sensibly, the gap explained by D&A and working capital, not aggressive accruals. Inventory at ~$4.7–4.8B / 209 days sits inside TI’s own 150–250-day target and is a deliberate cycle-readiness choice, not a hidden write-down.

But two QoE flags belong on the FCF line. First, reported “free cash flow” is materially flattered by CHIPS Act cash: FY25 CFO included a ~$335M ITC tax benefit (FY24: ~$588M), and TI received $555M of direct CHIPS proceeds in Q1-26 alone; the trailing FCF includes ~$965M of CHIPS incentives. A meaningful slice of the recent FCF recovery is government subsidy, not core operating cash. Second — more subtle — CHIPS incentives are booked as a reduction in the carrying value of the related fixed assets, which lowers depreciation over the assets’ lives (Fact — FY2025 10-K accounting-policy note). That will flatter reported gross margin for years, not just one-time cash. A clean read of normalized profitability strips both effects. The ITC steps up to 35% for assets placed in service after 12/31/2025, so this tailwind grows.

Balance sheet and returns: solid, but levered up at the wrong time

Net debt/EBITDA of ~1.1x (FY25 EBITDA ~$8.0B; total debt ~$14.15B against ~$4.9–5.1B cash) is comfortably investment-grade, and the debt is termed-out across a long, laddered profile (notes out to 2063). The discomfort is when the leverage was added: total debt rose from ~$8.2B (FY21) to $14.15B (FY25), and interest expense climbed, much of it issued in 2023–25 at 4.5–5.1% coupons — TI funded its dividend and capex through the trough at the highest rates of the cycle. ROE compressed from ~60% (FY21–22) to ~28–31% (FY24–25) and ROIC roughly halved, both reflecting the bloated asset base earning sub-trend returns until the fabs fill.

Verdict on Financial Quality: a genuinely high-quality, clean-accounting Analog franchise whose reported economics are currently depressed by a self-imposed capex/depreciation cycle and propped by subsidy. Through-cycle, economics improve with scale; right now the shareholder is mid-bet and underwater on FCF/share. High quality, worst-look phase.

7. Capital Allocation

Verdict up front: TI’s capital allocation has historically been a model of discipline — the explicit religion is “maximize long-term free-cash-flow per share,” and for two decades the company executed it ruthlessly: opportunistic buybacks, a 22-year dividend-growth streak, almost no empire-building M&A. The last six years strain that record. Management made a bold, leveraged bet — build 300mm capacity ahead of demand — that has so far reduced FCF/share rather than grown it; it sustained a dividend it could not cover from FCF by adding ~$6B of trough-priced debt; it largely stopped buying back stock during the very years the share price fell; and it has now done its first sizeable acquisition since 2011 at a rich multiple. None is indefensible, but together they mark a clear departure from FCF/share orthodoxy, and the payoff is still entirely prospective.

The capex supercycle: conviction or over-build? The defining decision is the six-year, ~$20B+ 300mm program. The strategic logic is real: 300mm carries ~40% lower unit cost than 200mm, owned fabs capture the margin internally (vs. foundry), and US/secure capacity is a genuine differentiator with industrial and automotive customers. CHIPS support — a 35% ITC plus up to $1.6B direct grants — improves after-subsidy economics. The boldness is building ahead of demand, deliberately accepting years of under-utilization and depressed margins to own dependable low-cost capacity for the 2030s. The risk is symmetric and large: if demand to fill RFAB2/LFAB/Sherman arrives on schedule, TI emerges with a structural cost moat and a powerful FCF re-rating; if analog demand stays soft, TI carries $2B+/yr of depreciation on half-empty fabs. So far the scoreboard reads over-build: FCF/share is below 2021. The 2026 capex step-down to $2–3B signals the build phase ending; the next two years reveal whether it was conviction or overreach.

Dividends: an admirable streak, now uncomfortably stretched. TI raised its dividend for the 22nd consecutive year, with dividends paid rising $3.89B (FY21) → $5.00B (FY25) — a genuine signal of commitment to returning cash. But coverage has gone sideways: FY25 dividends were 100% of net income and ~192% of FY25 free cash flow ($2.60B). For three straight years (FY23–FY25) the dividend exceeded FCF, the gap plugged by debt and cash. Interpretation: the dividend is not at imminent risk — through-cycle FCF easily covers it and the balance sheet has capacity — but maintaining the raise through the trough prioritized the streak over the FCF/share discipline TI preaches, financed with 4.5–5.1% debt. A values trade-off worth naming: the dividend religion temporarily outranked the FCF/share religion.

Buybacks: historically disciplined, but crowded out. TI’s buyback record is the strongest evidence for its capital-allocation IQ — it repurchases counter-cyclically and shrinks the count over time. But the recent pattern is mixed: buybacks were slashed from $3.62B (FY22) to $0.29B (FY23), then $0.93B (FY24) and $1.48B (FY25) — TI pulled back sharply precisely as the stock derated, because capex and the dividend consumed the cash. The entry prices were disciplined (FY25 ~$174/share, FY24 ~$198/share) but the size was small; net diluted share count is only modestly lower (~920M FY21 → ~913M FY25), recent buybacks roughly offsetting SBC dilution rather than meaningfully shrinking the base. A large multi-year authorization remains, so dry powder is there for when FCF normalizes. Interpretation: the discipline (buy low, never overpay) is intact, but the capacity to act was crowded out by the capex/dividend stack — TI bought less stock when it was cheapest, the opposite of what the FCF-per-share model prescribes with unlimited capital.

Silicon Labs: the first real deal since 2011 — and it’s expensive. On 2026-02-04 TI agreed to acquire Silicon Labs for ~$231/share, all-cash, ~$7.5B enterprise value (close expected 1H-2027), funded with cash plus ~$7B incremental debt — its first sizeable acquisition since National Semiconductor (2011). The rationale — embedded wireless connectivity to shore up the Embedded Processing segment that has structurally weakened (op profit down ~70% in two years) — is strategically coherent. On price, ~$7.5B EV against Silicon Labs’ ~$0.8B revenue is roughly 9–10x EV/sales for a thinner-margin business, and management’s framing (first-full-year EPS-accretive ex-deal-costs, >$450M annual synergies within ~three years from moving ~10–15 dies onto TI’s Lehi capacity, zero modeled revenue synergies, leverage-neutral 18–24 months post-close) is disciplined but rests on a low accretion bar. Interpretation: a defensible strategic fill-in but a richly priced one, layered on an already-stretched capital plan, and a signal that organic Embedded execution fell short. It deserves scrutiny, not a pass on management’s M&A track record.

Incentives and insider behavior: aligned on paper, net sellers in practice. The proxy (DEF 14A, filed 2026-03) ties pay to the right things — 1- and 3-year absolute and relative-to-semiconductor-peer performance on revenue, operating margin, FCF/share growth, and relative TSR, with the stated north star of “long-term growth of free cash flow per share.” No soft ESG metrics pollute the scorecard and SBC is low, so management is not diluting shareholders to pay itself. The discordant note: the pay-vs-performance table shows TI’s $100 invested grew to ~$122 vs. the S&P IT Index’s ~$258 over the window — management was paid through a period of severe relative underperformance. The insider tape reinforces caution: across ~188 Form 4s (2023–2026) there were zero open-market purchases and ~127 sales (~$361M), led by Chairman Templeton (~$189M), the CFO (~$52M), and CEO Ilan (~$14M). Most sales are routine exercise-and-sell of vesting equity, not conviction dumps — but the complete absence of open-market buying through a multi-year ~50% derating is a mild negative. Add the CFO transition (Julie Knecht succeeds Rafael Lizardi effective 2026-08-01; 8-K 2026-06-02), and the steward of TI’s famed capital-discipline framework is changing hands just as the bet must pay off.

Verdict on Capital Allocation: historically excellent and still rational at its core, but the last six years are a clear, deliberate deviation from FCF/share orthodoxy — a leveraged, build-ahead capex bet financed alongside a debt-supported dividend and shrunken buybacks, capped by a pricey acquisition. Whether this was visionary counter-cyclical investment or a discipline lapse will be settled by one question the filings cannot yet answer: will the new fabs fill? Management has earned the benefit of the doubt on intent; it has not yet earned it on outcome.

8. Changes and Headwinds — Last Two Years

The last ~24 months reshaped TI’s leadership, its capital profile, and — for the first time in over a decade — its M&A posture. Several changes strengthen the long-term setup; a cluster of China-related and execution risks cut the other way.

Leadership succession — CEO then CFO, in close sequence. Rich Templeton handed the CEO role to Haviv Ilan in 2023 (Templeton remains Chairman) — an internal, continuity-oriented transition. More consequential for the doctrine: long-tenured CFO Rafael Lizardi is being succeeded by Julie Knecht effective 2026-08-01 (Board action 2026-05-27; 8-K filed 2026-06-02). Lizardi was the public architect of TI’s FCF/share framework, the capex-intensity model, and the 100%-of-FCF-return commitment. Knecht is an internal promotion (also Chief Accounting Officer), arguing for continuity — but losing the doctrine’s chief evangelist mid-cycle, just as the capex supercycle rolls off and SLAB integration begins, is worth monitoring. Mildly thesis-neutral-to-negative on the margin.

The capex supercycle peaks and rolls off — the single biggest financial change. With Sherman SM1 in production and major shells complete, TI shifted into modular, demand-pulled equipping. Capex falls to $2–3B in 2026 from $4.6B in 2025, and the long-run model is now ~1.2x revenue growth, aided by fab-throughput gains. This is the mechanical driver of the FCF inflection (TTM FCF $4.4B vs. $1.7B a year ago). Strongly thesis-positive — the heavy-investment phase that depressed FCF is ending.

The offsetting headwind: depreciation drag. The same capex now hits the P&L as depreciation: $1.9B (2025) → $2.2–2.4B (2026) → higher in 2027. This is why gross margin (58% Q1-26) and operating margin (~37%) remain far below the FY22 peak even as revenue recovers, and TI’s 10-K warns these fixed costs “do not decline with reductions in customer demand.” Margins will be a multi-year recovery, gated by utilization.

Silicon Labs — first major M&A since 2011. ~$231/share all-cash, ~$7.5B EV, ~$7B incremental debt, closing 1H-2027; covered in the Capital Allocation section. Strategically coherent and financially disciplined in framing, but adds debt and integration/regulatory (including China) risk.

China — the concentrated headwind. ~50% of revenue ships into China and ~20% comes from China-HQ customers. The Sept-2025 MOFCOM anti-dumping probe into US analog ICs, domestic competitors taking low/mid-end share, and SLAB requiring China regulatory clearance to close make China the densest cluster of risk. TI’s 10-K explicitly names China “reshaping its domestic semiconductor industry through policy changes and investment” as a competitive threat.

CHIPS Act — the geopolitical tailwind. A 35% ITC (up from 25%) on US capex plus up to $1.6B direct funding ($630M received through 4Q25, including $555M for Sherman; ~$965M of CHIPS cash in TTM FCF). Materially subsidizes the onshore footprint and is a genuine offset to the China overhang — though it flatters reported FCF and should be normalized in valuation.

Verdict: net thesis-strengthening, with the strengthening concentrated in cash flow and the weakening concentrated in China. The capex roll-off and FCF inflection are the dominant favorable change; the SLAB deal is disciplined and strategically sound; CHIPS subsidizes the moat. Against that, the CFO transition removes the doctrine’s architect, depreciation caps near-term margins, and China (probe + share loss + SLAB approval) is a real, partly binary overhang. On balance the changes strengthen the long-term thesis but raise the near-term execution and geopolitical bar.

9. Risk Analysis

Top risks, in prose.

1. Cyclicality / a renewed inventory correction (H/H). The dominant risk and the one management itself most clearly fears. The 2024 trough was preceded by a “historic” inventory correction; 2025 then delivered a strong H1 that decelerated into a “false start” H2. Q1-2026 is again a strong H1, and management explicitly and repeatedly declined to call the second half. Because TI carries ~209 days of inventory and owns most of its manufacturing (high fixed costs), a demand stall would simultaneously compress revenue, force under-loading, and amplify the depreciation drag — a triple hit to margins. The +30% industrial print is precisely the kind of replenishment number that can reverse.

2. China — geopolitical and competitive, intertwined (M/H and M/M). TI’s single largest concentration and densest risk cluster. ~50% of revenue ships into China; the Sept-2025 MOFCOM anti-dumping probe is a live, partly binary threat that could impose duties or accelerate substitution toward domestic suppliers. Separately, Chinese competitors are taking low/mid-end share — a structural (not cyclical) erosion TI’s own 10-K acknowledges. The two interact: geopolitical pressure gives Chinese customers cover and incentive to design out US parts. SLAB’s required China approval adds a third China-contingent variable to the deal close.

3. Capacity over-build / under-utilization & the depreciation drag (M/H and H/M). TI made a deliberate, contrarian ~$20B bet to build 300mm capacity into a downturn. If demand is slower to arrive than modeled (the Marathon capital-cycle red flag of asset growth into weak returns), TI runs fabs below optimal loading while depreciation climbs to $2.4B in 2026 and higher in 2027. This is why margins remain ~11 points (gross) and ~14 points (operating) below the FY22 peak despite recovering revenue. The depreciation drag is near-certain; the severe under-utilization scenario depends on the demand cycle.

4. SLAB integration / overpayment (M/M). The ~$7.5B price implies a full multiple, and the accretion case rests on COGS synergies that don’t fully land until ~2030, plus zero modeled revenue synergies. Integration of a software-rich wireless business, the die-transfer execution to Lehi, and ~$7B of new debt all carry execution risk — though TI’s disciplined framing mitigates the empire-building concern.

10. Valuation Discussion

Framing: this section analyzes what the ~$282 price embeds — not whether to own it. No price target, no recommendation.

Where TXN trades — peer comps

At ~$282 (≈$256B market cap, ≈$265B EV on ~$9B net debt), TI carries a premium analog multiple. The trailing P/E of ~48x is not the right anchor: it sits on cyclically depressed earnings (FY25 EPS ~$5.50 vs. the FY22 peak of ~$9.41), so the trailing multiple is mechanically inflated. The forward P/E (~30x) and EV/EBITDA (~31x) are the cleaner read.

Peer multiple table (public market data, 2026-06-09; trailing P/E distorted by trough earnings across the group):

The honest read: TXN trades in line with ADI — its closest quality peer — on EV/EBITDA (31x vs. 33x) and EV/Sales (15x vs. 16x), at a meaningful premium to NXPI (16x forward, auto/industrial-cyclical, lower-margin) and to onsemi. ADI’s reported ROE (9.6%) is depressed by Maxim acquisition amortization, so TXN screens as the higher-return, owned-fab “best-of-breed” name and is priced accordingly. MPWR is the outlier richness of the group (49x forward) and not a fair comp for a $18B-revenue mature franchise. The trough-earnings P/E noise (MCHP 400x, ON 81x) underscores that this sector is being valued through a cyclical bottom on the expectation of normalization — TXN included.

Own-history percentile — the richest 2% of its own decade

The most striking valuation fact is internal, not relative. TXN’s own-history valuation percentiles (2026-06-09) place it at roughly the 98th composite percentile of its trailing ~10-year range:

On every lens that nets out cyclicality more than P/E does (P/S, P/B), the stock is more expensive than at virtually any point in the last decade. Part is deliberately explained — earnings and FCF are trough-depressed by the capex/depreciation supercycle, so price-to-depressed-fundamentals overstates richness. But P/S strips earnings volatility out, and even on sales TXN sits at its 99th percentile. The market is paying an all-time-high multiple of revenue for a company whose revenue has barely grown off the 2022 peak. That is the core valuation tension.

Embedded expectations — what must be true at $282

Reverse the price into the implied fundamentals:

• Versus management’s own target. Management calls ~$8 FCF/share (~$7.3B) for 2026 “highly probable” if revenue grows mid-to-high single digits. At ~$282, the stock trades at ~35x that FCF target — a ~2.9% forward FCF yield. For a mature, cyclical, capital-intensive analog franchise, a sub-3% FCF yield is a growth-and-normalization multiple, not a value multiple.
• Steady-state reverse-DCF. To support a ~$265B EV at a 4% exit FCF yield, the market needs roughly $10.6B of steady-state FCF; at 4.5%, ~$11.9B. Against management’s own framework ($20B revenue → $8–9B FCF; $22B → $9–10B FCF), reaching ~$11B FCF requires revenue near $24–26B at ~45–50% FCF conversion — roughly a 25–30% revenue lift from the FY25 base plus full margin recovery plus capex at its long-run floor, all simultaneously.
• Two-stage check. Discounting ~$8B of FCF growing ~10%/yr for a decade to a 3% terminal at a 10% cost of equity lands roughly at today’s price. So the embedded expectation is explicitly a return to FY21–22-vintage FCF generation, sustained, with continued capex roll-off and mid-single-digit-plus revenue CAGR. Each is plausible; requiring all three together is the underwriting risk.

What the market is pricing correctly: that capex is rolling off, that depreciation is the near-term earnings headwind and transient, and that the owned-fab/300mm cost base is a durable structural advantage. What looks aggressive: the speed and completeness of margin normalization, and crediting the full FCF recovery into a sub-3% yield while depreciation is still climbing and gross margin sits at 57% — ~12 points below the peak.

The normalized-margin debate — the swing variable

Every scenario hinges on where gross margin settles. Bull view: as the underutilized new 300mm fabs fill, gross margin re-expands toward the low-60s% and operating margin toward the 40s%, restoring FCF toward $11–12B; the CHIPS ITC reduces PP&E carrying value, lowering future depreciation and flattering the reported line. Bear view: owned-fab depreciation establishes a permanently lower ~57–60% gross-margin base versus the fabless-light prior peak — the 68.8% high was a capacity-light artifact that won’t recur — so “normalized” is ~57–60% GM / low-30s% OpM, and the ~$8 FCF/share target is closer to a ceiling than a waypoint. This single assumption moves the valuation by tens of billions, which is why the scenario spread is so wide.

Scenario analysis (2028–2030 normalized; ~909M shares, ~$9B net debt)

The asymmetry is informative. The base case roughly validates today’s price — meaning at $282 the market already underwrites a successful, fairly complete cyclical and margin recovery. The bull case (+~35%) requires the optimistic margin path and a revenue level (~$25B+) the company has never reached. The bear case (−~45%) is not a business-failure scenario at all — it merely assumes the lower normalized-margin base and no revenue breakout, on a de-rated (but still reasonable) ~22x multiple. The skew is unfavorable: limited room above for the base case, a long way down if normalization disappoints or the multiple compresses toward the group’s NXPI-like end.

Dividend support

The ~2.0% dividend yield (~$5.70/share) and TI’s long, credible record of per-share return growth provide a valuation floor and a reason patient capital holds through the cycle. The constraint: the payout ratio is currently ~80–95% of depressed earnings, so dividend growth is throttled until FCF recovers — the dividend supports the floor but cannot, near-term, drive the case. A cushion, not a catalyst.

Net: at ~$282, TXN is priced at the top ~2% of its own decade and roughly in line with ADI, embedding a near-complete margin/FCF normalization (~$8–9B+ FCF) at a sub-3% forward FCF yield. The market is pricing the capex roll-off and the fab cost advantage correctly; it is pricing the magnitude and certainty of margin recovery generously while depreciation is still rising. The base case justifies the price; the disconfirming risk is the durable-margin debate and China.

11. Variant Perception

Consensus view. The market broadly treats TI as a best-in-class, wide-moat analog franchise whose multi-year capex investment is now ending, inflecting free cash flow sharply higher (TTM FCF $4.4B vs. $1.7B a year ago; >$8 FCF/share “highly probable” for 2026). The cyclical recovery is underway (eight straight sequential growth quarters), pricing has stabilized, data center is a new secular leg, and the 100%-FCF-return + 22-year dividend-growth machine is intact. Consensus underwrites: capex roll-off → FCF inflection → re-rating, with a structurally advantaged manufacturing/portfolio moat.

Strongest bull case. TI is a genuine economies-of-scale + cost-advantage business: owned 300mm/28nm manufacturing at structurally lower cost than foundry-reliant peers, ~95%-internal wafers by 2030, the industry’s broadest analog/embedded catalog (the “alpha-socket → attach” cross-sell), and geopolitically-dependable US capacity subsidized by a 35% ITC. The heavy-capex phase is over precisely as the cycle turns up, so incremental revenue drops through at 75–85% with minimal new investment. If industrial (still ~15% below peak) recovers toward and past prior peaks, data center compounds, auto content keeps rising, pricing turns up in 2H-2026, and SLAB adds a connectivity leg, FCF/share compounds for years toward the $9–10B-at-$22B-revenue framework — a high-quality compounder whose earnings power the market is anchoring to depressed trough margins.

Strongest bear case. The recovery is flattered by trough math and a partly-spent capital cycle. Industrial +30% and reported +19% are replenishment off a deep base; 2025 already proved a “false start,” and 2026’s H2 could repeat it. China — ~50% of shipped revenue — faces a live anti-dumping probe and steady domestic share loss in low/mid-end analog, a structural erosion of TI’s volume base. The ~$20B capacity bet risks under-utilization with depreciation climbing through 2027, capping margins far below the FY22 peak for years. SLAB adds ~$7B debt and integration risk for a business that may not be GAAP-accretive near-term, while the CFO who authored the capital-return doctrine departs. On this read, “normalized” FCF is lower and later than bulls assume, and the stock pays a premium multiple for a cyclical recovery dressed as secular compounding.

The 3–5 assumptions that matter most:

1. Is the 2026 recovery durable, or a second “false start”? The entire FCF-inflection thesis rests on H2-2026 not decelerating the way H2-2025 did.
2. Does data center stay a secular growth leg, or normalize to a competitive, slower-growing ~10%? The +90% rate is the most-quoted bull data point and the least sustainable.
3. How much low/mid-end China share is permanently lost — and does the MOFCOM probe escalate? This sets the structural ceiling on TI’s largest geography.
4. Where do margins normalize given the depreciation schedule? Whether gross margin re-approaches the high-60s or stalls in the low-60s swings normalized FCF materially.
5. Is SLAB accretive on the promised timeline, and does the ~$7B debt + 100%-return pledge prove compatible through the cycle?

What would falsify each side:

• Falsifies the bull: Q2/Q3-2026 sequential growth stalls or reverses; MOFCOM imposes duties or China customers visibly design out TI parts; depreciation-driven margin recovery stalls in the low-60s gross despite revenue gains; SLAB requires guidance cuts or fails to close on China approval.
• Falsifies the bear: H2-2026 sustains double-digit YoY growth with pricing increases landing; industrial pushes past its 2022 peak; data center holds a high growth rate as application-specific sockets ramp into 2027; gross margin marches back toward the high-60s as utilization rises and capex stays at the $2–3B floor.

12. Fact vs. Interpretation

13. Open Questions

1. H2-2026 sustainability. Does the strong first half hold, or is it a second consecutive “false start”? This single question gates the FCF-inflection thesis. Watch: Q2 (Jul-2026) and Q3 (Oct-2026) sequential prints, lead times, order linearity.
2. Normalized gross margin. Where does GM settle once the new fabs fill — high-60s (bull) or low-60s/high-50s (bear)? Management has not committed to a normalized target. Watch: incremental gross-margin fall-through (guided 75–85% ex-depreciation) vs. the depreciation schedule.
3. China share and the MOFCOM probe. How much low/mid-end industrial revenue shipped into China is permanently lost, and does the anti-dumping probe escalate to duties or design-outs? Watch: probe rulings; China-region revenue trend; SLAB China approval as a signal.
4. Data-center durability. Does the ~$450M/quarter run-rate hold a high growth rate as application-specific (VRM/GaN) sockets ramp into 2027, or normalize toward competitive single-digit growth?
5. CFO transition. Does Julie Knecht preserve the FCF/share doctrine and 100%-return discipline, or does the framework drift post-Lizardi?
6. SLAB accretion and close. Is the deal GAAP-accretive on the promised timeline, do the >$450M synergies land, and does it clear regulators (including China) on schedule in 1H-2027?

14. What Must Be True

For the bull case to be right (constructive on a 3–5 year view):

• Industrial recovers toward and past its 2022 peak (still ~15% below) and the broader 2026 upcycle does not stall in H2 the way 2025 did.
• Gross margin normalizes back toward the low-to-mid 60s% as utilization rises and the CHIPS ITC suppresses reported depreciation — restoring FCF toward $9–12B.
• Data center holds a high (even if decelerating) growth rate, and China low/mid-end share loss stays a slow leak rather than an accelerating breach.
• Falsification test: two consecutive quarters of flat-to-negative sequential revenue in 2026–27, or gross margin stuck below ~60% a year after revenue exceeds $19B, or a MOFCOM duty/design-out event — any one breaks the “FCF compounding / margin recovery” core.

For the bear case to be right (the price overstates normalized economics):

• “Normalized” gross margin proves to be a permanently lower ~57–60% owned-fab base, capping normalized FCF nearer $6B.
• Revenue fails to break decisively above the $20B 2022 peak as Chinese substitution and a maturing auto/PE base offset industrial/data-center gains.
• The multiple de-rates from ~30x forward / ~14x sales toward the analog group’s lower end as the “inflection” proves partial.
• Falsification test: TI prints two consecutive quarters of >60% gross margin on revenue above $19B with capex held at the $2–3B floor, demonstrating the margin step-down was cyclical, not structural — which would validate the bull’s normalized-FCF math and break the bear.

15. Source Appendix

See the Source Appendix (Appendix B below) for the full primary-source list. Principal sources: TI FY2025 Form 10-K (filed 2026-02-06) and prior 10-Ks; Q1-2026 Form 10-Q; Q1-2026 earnings call (2026-04-22); Silicon Labs M&A call (2026-02-04); Capital Management / Special call (2026-02-24); Shareholder/Analyst call (2026-04-16); Bernstein Strategic Decisions Conference presentation (2026-05-28); DEF 14A (2026-03); 8-K material-event filings (incl. CFO transition 2026-06-02); SEC EDGAR XBRL financial concepts (CIK 0000097476); and public peer/industry data.

APPENDIX A — Standard Diligence Questionnaire — Texas Instruments (NASDAQ: TXN)

Supplemental to the main note. Fact / Interpretation / Assumption labels applied where it matters. As-of 2026-06-10, price ~$282.

General

What thoughtful questions have other investors asked about this company? The recurring sell-side and buy-side questions (from the Q1-2026, Q4-2025, and Capital Management calls) cluster on five themes: (1) Is the 2026 recovery durable or another “false start” like H2-2025? (2) Where does gross margin normalize given the depreciation schedule — is 57% a trough or a new base? (3) Free-cash-flow-per-share trajectory — when does the capex roll-off translate to the $8+/share the framework promises, and what fills the new fabs? (4) Data-center power durability — is +90% YoY a one-off supply-gap fill or a secular leg? (5) Pricing — does TI follow the market up in 2H-2026, and what share of revenue is on annual price contracts? Stacy Rasgon (Bernstein) pointedly questioned whether Silicon Labs is GAAP-accretive at all near-term. (Interpretation, from earnings/conference transcripts.)

Cyclicality & Earnings Nature

Are earnings at a cyclical high or low? Low-to-recovering. FY24 EPS (~$4.99) was a cyclical trough; FY25 (~$5.50) and TTM (~$5.85) are recovering but remain far below the FY22 peak (~$9.41). Margins are trough-depressed by under-utilization plus new-fab depreciation. (Fact — EDGAR.)

Driven by external environment or internal actions? Both, distinctly. Revenue is driven by the external analog/industrial cycle; margins are depressed by an internal choice — the ~$20B capex supercycle that loaded depreciation ahead of demand. The earnings trough is therefore deeper than the demand trough alone would imply. (Interpretation.)

How stable are revenues? Cyclical, not stable — 22% peak-to-trough (FY22→FY24). But the base is diversified (100,000 customers, 80,000 products, ~half of revenue outside the top-50 customers) and the underlying design-in positions are long-lived (10–20+ year part lives), so the business does not lose customers in a downturn — it loses volume that returns. (Fact/Interpretation — 10-K.)

Outlook for products/services? Secular content growth in Industrial (automation, electrification), Automotive (more silicon per vehicle), and Data Center (AI power). The portfolio has been deliberately shifted toward these (Industrial+Auto+Data Center ~75% of revenue, up from 43% in 2013). (Fact — 2026-02-24 call.)

How big will this market be? Analog TAM ~$93.7B (2025), growing mid-single-digits secularly, plus embedded. TI is #1 at ~19% share. Global, with ~50% of TI revenue shipped into China. (Fact — industry aggregators; 10-K.)

Business Quality & Competitive Moat

Is the industry getting more or less competitive? More, specifically at the low/mid-end, where Chinese domestic players (SG Micro, 3Peak, Novosense) are taking structural share, and in data-center power (Monolithic Power, Infineon, ADI). The high-performance, design-in-captive core remains oligopolistic. (Interpretation — 10-K + competitor data.)

How profitable is the business (ROIC, ROE)? ROE ~31% (FY25), down from ~60% (FY21–22) as the asset base ballooned; through-cycle ROIC is high but currently depressed. Analog segment operating margin 38.6%; consolidated 34.1% at trough. (Fact — EDGAR/10-K.)

How profitable is the industry — competitors, barriers? The cost leader (TI) and the precision leader (ADI) earn 35%+ operating margins; the cyclical IDMs (Infineon, NXP, ST, onsemi) earn 17–28%. Barriers: process/packaging know-how, multi-year auto/industrial qualification, catalog breadth, and 300mm manufacturing scale — real but lower than leading-edge logic. (Fact.)

Can the business be easily understood? Yes — a low-obsolescence catalog of analog/embedded chips sold to industrial/auto/consumer OEMs, run for FCF/share. Among the more transparent semiconductor models. (Interpretation.)

Can it be undermined by foreign low-cost labor? Not labor — but by state-subsidized foreign capacity: Chinese domestic analog makers attacking the commodity tail with lower prices and a localization mandate. The threat is policy/capital, not wages. (Interpretation.)

Do brands matter? Modestly — TI’s reputation for reliability, supply dependability, and breadth matters in design-in decisions, but the moat is cost + captivity + scale, not consumer brand. (Interpretation.)

Nature of competition? Design-in competition (winning the socket years ahead of production) plus supply reliability and price. TI competes on breadth + cost + dependable supply rather than bleeding-edge performance (ADI’s lane). (Interpretation.)

Customers’ switching costs? High in auto/industrial/embedded (re-qualification, functional-safety re-certification, firmware reuse, decade-long platforms); low in personal electronics and commodity catalog parts. (Fact/Interpretation — 10-K.)

Financial Condition & Balance Sheet

Assets not fully recognized on the balance sheet? The 80,000-product catalog and accumulated analog/process IP are internally developed and largely not capitalized — a real off-balance-sheet asset. Also: largely-depreciated legacy 200/300mm capacity carried at low book value but still productive. (Interpretation.)

Off-balance-sheet liabilities? Nothing material flagged; purchase/construction commitments for the fab build-out and operating leases are disclosed in the 10-K. The ~$7B of incremental SLAB-related debt is contracted but the deal has not closed. (Fact — 10-K; M&A call.)

How conservative is the accounting? Very. Pure GAAP reporting (no non-GAAP addback parade), low SBC (~2.4% of revenue), inventory inside the stated 150–250-day band. The one nuance: CHIPS incentives reduce PP&E carrying value, lowering future reported depreciation (a reported-margin flatter, fully disclosed). (Fact — 10-K.)

How CapEx-hungry is the business? Very, by choice, right now — ~$5B/yr at the 2021–24 peak (~30% of revenue). Long-run guide is ~1.2x the revenue-growth rate (e.g., 5% growth → ~6% of revenue capex), i.e., normalizing to ~$2–3B. An IDM is structurally more capex-hungry than a fabless peer, the trade-off for the cost advantage. (Fact — Q1-26 call.)

Capital Allocation & Management

How much FCF, and how is it used? FY25 FCF ~$2.6B (trough, depressed by capex); TTM ~$4.4B and inflecting. Policy: return 100% of trailing FCF to owners via dividends + buybacks. Currently dividend-heavy (~$5B/yr); buybacks were cut to fund capex. (Fact — 10-K, calls.)

Significant acquisitions recently? Yes — Silicon Labs (~$231/share, ~$7.5B EV, all-cash, closing 1H-2027) — first sizeable deal since National Semiconductor (2011). Strategic (embedded wireless connectivity) but richly priced (~9–10x EV/sales). (Fact — 2026-02-04 M&A call.)

Buying back shares? Yes, but minimally during the capex trough ($0.29B–$1.48B/yr FY23–25, after $3.62B in FY22). Net share count roughly flat — recent buybacks mostly offset SBC. A large authorization remains for when FCF normalizes. (Fact — EDGAR.)

Issuing large amounts of stock to insiders? No — SBC is low (~$419M, ~2.4% of revenue), among the lowest in large-cap semis. (Fact — 10-K.)

Compensation policy of directors/management? Pay tied to 1- and 3-year absolute and peer-relative revenue, operating margin, FCF/share growth, and relative TSR — no soft ESG metrics. Conceptually well-aligned. Caveat: management was paid through a period of severe relative underperformance (TI total return ~$122 vs. S&P IT ~$258 per $100 over the proxy window). (Fact — DEF 14A 2026-03.)

Motivations of management? The stated north star is “long-term growth of free cash flow per share.” Insider behavior: across ~188 Form 4s (2023–26), zero open-market purchases and ~127 sales (~$361M, mostly routine exercise-and-sell). The absence of conviction buying through a ~50% derating is a mild negative. CFO transition (Lizardi→Knecht, eff. 2026-08-01) changes the framework’s chief steward. (Fact — Form 4 corpus; 8-K 2026-06-02.)

Valuation & Market Data

ADR, MLP, or K-1 issuer? No — a US-domiciled C-corp (Delaware), common stock on NASDAQ. Standard 1099 dividend treatment. (Fact.)

Dividend policy? ~$5.70/share annualized, ~2.0% yield, 22 consecutive years of increases; intends to return 100% of trailing FCF (dividends + buybacks) over time. Payout currently high (~80–95% of depressed EPS, >100% of trough FCF), throttling near-term dividend growth. (Fact — 10-K, 8-Ks.)

How profitable is the business? Highly — 57% gross / 34% operating margin at the trough; ~31% ROE. (Fact.)

Is net income diverging from cash from operations? Sensibly aligned — FY25 NI $5.0B vs. CFO $7.15B, the gap explained by D&A and working capital, not aggressive accruals. The more important divergence is CFO vs. FCF, where elevated capex (now rolling off) is the wedge. (Fact — EDGAR.)

Risks & Downside

What factors would cause the stock to decline? A 2H-2026 demand stall (a “false-start” repeat); a China escalation (MOFCOM duties or accelerated design-outs); evidence that normalized gross margin is stuck below ~60%; SLAB integration/close problems; or simple multiple compression from a ~98th-percentile-of-own-history starting point. (Interpretation.)

Risk of a catastrophic loss? Low. Investment-grade balance sheet (~1.1x net debt/EBITDA), diversified revenue, durable design-in base, and a real cost advantage. The risk is valuation/return, not solvency. (Interpretation.)

Chance of a total loss? Negligible — a profitable, cash-generative, investment-grade market leader. The realistic downside is a ~40% de-rating to the bear-case zone, not impairment of the enterprise. (Interpretation.)

Recent News & Events

Has the business environment changed recently? Yes: (1) cyclical recovery underway (8 straight sequential growth quarters); (2) capex supercycle peaking and rolling off → FCF inflecting; (3) China anti-dumping probe (Sept-2025) + rising domestic competition; (4) Silicon Labs acquisition announced; (5) CFO transition (Aug-2026); (6) CHIPS Act ITC stepped up to 35%. The news flow (2026-06-09/10) was otherwise quiet — a Wells Fargo price-target raise to $300 (equal-weight) and a new battery-monitor product launch, neither thesis-changing. (Fact — 8-Ks, calls, press.)

Significant acquisitions? Silicon Labs (above). Change in accounting policies? None material; CHIPS-incentive accounting (PP&E carrying-value reduction) is disclosed. Recent changes — new markets, facilities, management? New 300mm fabs (Sherman SM1 in production, Lehi, Richardson); data-center end-market broken out in FY2025 resegmentation; CEO (2023) and CFO (2026) successions, both internal. (Fact.)

APPENDIX B — Source Appendix — Texas Instruments (NASDAQ: TXN)

Primary sources prioritized. Accessed 2026-06-10 unless noted. Internal/Drive cross-reads labeled. All financial figures reconciled to SEC filings (EDGAR CIK 0000097476) where possible.

Primary — SEC filings (EDGAR, CIK 0000097476)

Primary — management calls & presentations

Primary — quantitative data tools

Secondary — industry & competitive context

All material claims in this note are grounded in the primary public sources above.