Texas Instruments Incorporated (TXN) - Deep Dive Research Report

Prepared 2026-06-20. Six most recent earnings calls used: Q4 2024 (Jan 23 2025), Q1 2025 (Apr 23 2025), Q2 2025 (Jul 22 2025), Q3 2025 (Oct 21 2025), Q4 2025 (Jan 27 2026), Q1 2026 (Apr 22 2026).

1. What the company does

Texas Instruments makes the small, unglamorous chips that almost every electronic device needs but almost nobody asks about by name. If a product has a circuit board inside it - a car, a factory robot, a washing machine, a server rack, a hearing aid, a satellite - there is a high chance several TI chips are sitting on that board doing essential plumbing work: converting one voltage to another, cleaning up a sensor signal, charging a battery, driving a motor, or running a tiny embedded control loop. TI is the largest maker in the world of this category of chip, called analog and embedded semiconductors.

The distinction matters. Most of the semiconductor headlines (Nvidia, TSMC, the AI build-out) are about digital logic chips, which process ones and zeros and live or die on being made at the smallest, most advanced process node. TI plays a different game. Analog chips deal with the messy continuous physical world: heat, light, sound, pressure, voltage, current. They are not made on bleeding-edge nodes; many of TI's are made on geometries that are years or decades old. What makes them valuable is not transistor density but design know-how, reliability over a 10-to-20-year product life, and the sheer breadth of the catalogue. TI sells roughly 80,000 distinct products to more than 100,000 customers. No single chip is large; the business is the aggregation of tens of thousands of low-priced, high-margin, long-lived parts that customers design in once and buy for a decade.

TI was founded in 1930 (originally Geophysical Service, an oil-exploration outfit) and became a semiconductor pioneer - Jack Kilby invented the integrated circuit at TI in 1958. For decades it was a sprawling conglomerate: defence electronics, calculators, memory chips, digital signal processors, even laptops. The company that exists today is the product of two deliberate decisions. First, in the late 1990s and 2000s it exited commodity businesses (it sold its DRAM memory business to Micron, divested defence) and concentrated on analog and embedded. Second, the 2011 acquisition of National Semiconductor for about $6.5 billion roughly doubled its analog product breadth and sales force reach. The result is a company that today derives the overwhelming majority of revenue from analog and embedded chips sold into industrial and automotive customers.

The current strategic obsession is manufacturing. Analog chips can be made on either 200mm (8-inch) or 300mm (12-inch) silicon wafers. A 300mm wafer yields roughly 2.3x more chips than a 200mm wafer but costs only modestly more to process, so the unpadded die cost on 300mm runs roughly 40% lower. TI is the only large analog maker betting its balance sheet on owning that 300mm cost advantage at scale, in its own fabs, much of it on US soil. The bet captures the essence of the business:

"Once complete, [the elevated CapEx cycle] will uniquely position TI to deliver dependable, low-cost 300mm capacity at scale to meet customer demand." (Q4 2024 concall, Jan 23 2025)

A concrete example of what TI does for a customer: an automaker designing a new electric vehicle needs to manage the flow of power from a 400-volt or 800-volt battery down to the dozens of voltages every subsystem needs (the infotainment screen, the airbag controller, the headlights, the battery-management system that watches each cell). It needs to sense temperature and current everywhere, isolate high-voltage domains from low-voltage ones so a fault cannot electrocute a passenger, and drive the motors. A single EV can carry several hundred to over a thousand TI chips doing exactly this work. The automaker's engineers select those parts years before the car ships, qualify them through brutal automotive reliability testing, and then buy them for the entire production life of the model. TI's job is to have the right part in the catalogue, prove it will survive 15 years under a hot hood, and supply it reliably and cheaply for a decade.

2. Business segments

TI reports in three segments: Analog, Embedded Processing, and Other. The first two are the business; Other is a collection of legacy and miscellaneous lines.

Analog (~76% of revenue)

Analog is the heart of TI. It comprises two big product families. Power products take electricity and manage it: converting voltages up or down, regulating supply, charging batteries, driving motors and LEDs, and managing the power inside everything from a phone to a server rack. Signal Chain products are the bridge between the physical world and digital processors: amplifiers, data converters (chips that turn a real-world analog signal into digital numbers and back), interface and isolation chips, and sensors. Together they are sold into every end market TI serves.

The core capability here is design experience and catalogue breadth that took decades to assemble. An analog part is hard to design well: it has to behave predictably across temperature, voltage, and manufacturing variation, and small layout choices change performance in ways that are not fully simulatable. TI employs thousands of analog engineers, many with very long tenures, and the institutional knowledge of how to design a part that yields well and survives field conditions is genuinely hard to replicate. Layered on top is the manufacturing cost advantage: TI makes most of its own analog chips in its own fabs, increasingly on 300mm, which lets it undercut competitors who outsource to foundries or run older 200mm lines.

Analog exists as the dominant segment because it is the most defensible and highest-margin part of the company, and because the economics reward exactly TI's strengths - breadth, long product life, and low unit cost. Within the segment TI competes against Analog Devices (especially in high-performance signal chain), Infineon and STMicroelectronics (especially in power), and a long tail of others. TI generally wins on power management breadth and cost; it is less dominant in the highest-performance precision signal-chain niches where Analog Devices is strong. Management treats Analog as both the margin engine and the primary beneficiary of the 300mm investment. In Q1 2026 Analog revenue grew 22% year over year (Q1 2026 concall, Apr 22 2026).

Embedded Processing (~14% of revenue)

Embedded Processing is TI's microcontroller and processor business: small brains that sit on a board and run dedicated control software. This includes microcontrollers (MCUs) that run a single function reliably (a motor controller, a sensor hub, a body-control module in a car) and larger embedded processors for more compute-intensive edge tasks. These are not the application processors in your phone; they are the workhorse controllers embedded inside industrial and automotive equipment.

The core capability is a combination of processor design, a deep software and tools ecosystem, and the same automotive-grade reliability discipline as Analog. Customers who design a TI microcontroller into a product also adopt TI's development tools, libraries, and code; that software investment is a switching cost in its own right. Embedded exists as a separate segment because the design skills, the software ecosystem, and the competitive set differ from pure analog: here TI competes more directly with Microchip, NXP, Infineon (post-Cypress), STMicroelectronics, and Renesas, all of which have strong MCU franchises.

Embedded has historically been the weaker, lower-margin sibling and went through a multi-year repositioning - TI consolidated it onto its own internal 300mm manufacturing and trimmed less-strategic product lines. Management frames it as a growth segment being rebuilt to TI's manufacturing and margin standard rather than a cash cow. Embedded grew 12% year over year in Q1 2026, lagging Analog's 22% (Q1 2026 concall, Apr 22 2026).

Other (low single-digit % of revenue)

Other is the leftover bucket: DLP products (the digital micromirror display technology used in some projectors and automotive displays), TI's famous calculators (the graphing calculators still standard in US classrooms), custom ASIC work, and royalties. It is small, stable, and strategically peripheral. Management spends almost no airtime on it.

3. Products and business detail

TI's catalogue is its product. Roughly 80,000 orderable parts span a handful of functional families:

• Power management ICs: DC-DC converters, buck/boost regulators, battery chargers and battery-management systems, power-management ICs (PMICs), gate drivers, and motor drivers. These are the highest-volume, broadest part of the business. A modern car, server, or factory tool needs dozens of distinct voltage rails and TI supplies the chips that create and police each one.
• Signal-chain products: operational amplifiers, data converters (ADCs and DACs), interface chips (the parts that let chips talk to each other over standard buses), isolation chips (which pass a signal across a high-voltage barrier safely), clocks, and sensors (temperature, current, position). These are the eyes, ears, and translators between the analog world and the digital one.
• Microcontrollers and embedded processors: low-power MCUs for control tasks, plus more capable processors for edge compute, increasingly including the wireless-connectivity MCUs used in connected industrial and consumer devices.
• Specialty products: mmWave radar sensors (used in automotive driver-assistance and industrial sensing), DLP micromirror chips, and the calculator line.

The technical moat is partly in qualification, not just design. Automotive and industrial customers require parts certified to standards (AEC-Q100 for automotive ICs, functional-safety standards like ISO 26262) and demand a decade-plus of guaranteed supply. Getting a part qualified, proving long-term reliability, and committing to long supply lifetimes is a barrier most low-cost entrants cannot clear.

The defining operational story is manufacturing internalisation on 300mm. TI runs front-end wafer fabs across three principal sites in the current expansion: Richardson, Texas (RFAB1 and RFAB2, the latter the world's first 300mm analog fab built specifically for that purpose); Lehi, Utah (LFAB, acquired from Micron in 2021 for about $1.5 billion - it was Micron's 3D XPoint memory fab, repurposed for analog); and a large new multi-fab campus in Sherman, Texas (the SM1 / SM2 / SM3 / SM4 plan, the first of which began production in 2026). TI also runs assembly-and-test operations and is increasingly internalising back-end packaging as well as front-end wafers.

The strategic arithmetic: TI is moving from outsourced and 200mm production toward owning the bulk of its wafers in-house on 300mm. By management's own framing it grew the 300mm share of analog wafers from roughly 40% toward 60% during 2024 and targets the great majority of wafers being both internal and on 300mm by the end of the decade. The point is structural cost: a 300mm analog wafer yields far more chips than a 200mm one at little extra processing cost, so as the mix shifts TI's unit cost falls and its gross margin floor rises. Much of this US-based capacity is partly funded by the CHIPS Act: TI has a direct-funding agreement worth up to $1.6 billion, of which it had drawn about $630 million by Q1 2026 ($555 million arriving that quarter), with the rest tied to hitting construction and production milestones (Q1 2026 concall, Apr 22 2026).

Geographically TI sells worldwide. Its largest end-market exposure is industrial and automotive, sold into customers across the US, Europe, Japan, China, and the rest of Asia. China is both a large market and a competitive battleground, with a rising domestic Chinese analog industry. TI manufactures predominantly in the US (with assembly/test also in Asia), which it increasingly markets as a geopolitical-reliability feature: in a world worried about supply-chain security, a chip supplier that controls its own US fabs can promise continuity that fabless or foreign-fab competitors cannot.

4. Customers

TI's customer base is unusually wide and unusually un-concentrated. More than 100,000 customers buy from it, from the largest automakers and industrial conglomerates down to tiny hardware startups ordering parts off ti.com. The end-market mix has shifted deliberately over fifteen years: industrial, automotive, and data center together made up about 75% of 2025 revenue, up from roughly 43% in 2013 (Q4 2025 concall, Jan 27 2026). Industrial is the single largest end market (around 40% of revenue), automotive next (around 30%), with personal electronics, communications equipment, enterprise systems, and data center making up the rest.

Who actually decides to buy is an engineer, not a procurement officer. The buying decision happens at design-in: a hardware engineer at an automaker, an industrial-equipment maker, or a contract manufacturer selects a specific TI part while designing a circuit board, often years before volume production. The criteria are the part's specification fit, its reliability data, the quality of the supporting documentation and reference designs, availability of evaluation boards and software, and confidence that TI will still supply the part in a decade. Price matters but is rarely the first filter for a 30-cent part on a $30,000 machine. The sales cycle is long - design-in to volume can run two to four years in automotive and industrial - but the payoff is durable.

Why customers choose TI: catalogue breadth (one supplier can fill most of a board), the depth of free design support and reference designs, automotive/industrial qualification and long-life supply commitments, and increasingly supply assurance from TI's owned US capacity. The switching costs are real but quiet. Once a part is designed into a board and qualified, replacing it means redesigning, re-qualifying, and re-certifying - expensive and risky for a product that may ship for a decade. The lock-in is not contractual; it is the cost and risk of redesign, multiplied by tens of thousands of long-lived designs. Embedded adds a software lock-in on top, because the customer's firmware is written against TI's microcontroller and tools.

Concentration is low and that is a feature. No single customer dominates revenue; the long tail of industrial customers in particular is fragmented across thousands of accounts. This makes revenue resilient to any one customer leaving but also means TI's fortunes track broad industrial and automotive production rather than any single program. Contracts are mostly not long-term take-or-pay; the business runs on a flow of design wins converting to ongoing orders, plus a large catalogue/distribution channel and direct online sales. That structure makes revenue cyclical (it tracks industrial and auto build rates) but, across a cycle, highly predictable because of the breadth and stickiness of the installed base.

5. Competitive landscape

The analog and embedded market is an oligopoly of large, profitable incumbents plus a fragmented tail and a rising Chinese cohort. The top five suppliers - TI, Analog Devices, STMicroelectronics, Infineon, and NXP - together take roughly half the analog market. Each plays a slightly different game.

Analog Devices (ADI) is TI's closest peer and its main rival in signal chain, where ADI is generally regarded as the leader in the highest-performance precision converters and amplifiers, strengthened by its Linear Technology and Maxim acquisitions. TI competes with ADI more on breadth and cost than on the absolute top of the performance curve.

Infineon and STMicroelectronics are the European powers, strongest in power semiconductors, automotive, and (for Infineon and ST) silicon carbide - the wide-bandgap material increasingly used in EV powertrains, where TI is comparatively less exposed. Infineon's vertical integration in silicon carbide substrates is a real edge in that specific niche.

NXP is automotive- and IoT-heavy, strong in automotive processors, connectivity, and security. Microchip and Renesas are microcontroller-centric and compete with TI's Embedded segment more than its Analog segment. onsemi is strong in power and image sensing for automotive. Below the majors sits a long tail and, increasingly, a wave of Chinese analog companies competing aggressively on price in commodity power parts inside China - a structural competitive shift TI watches closely.

TI wins on three things: the broadest catalogue (a customer can single-source most of a board), the lowest structural manufacturing cost as the 300mm internal mix rises, and supply reliability from owned US capacity. It is more exposed where the competition is about a specific high-performance niche (ADI in precision signal chain) or a specific materials technology (silicon carbide power, where Infineon and ST lead), and where Chinese domestic competition is commoditising low-end power management. The barriers to entry are high but not insurmountable: building an 80,000-part catalogue, a global field-engineering force, automotive qualifications, and owned 300mm fabs would take a new entrant a decade and tens of billions of dollars. That is exactly why no Western new entrant has appeared - the threat instead comes from subsidised Chinese incumbents climbing the value chain.

Market caps are peer-size references only, approximate, as of June 2026.

6. Industry

Demand for analog and embedded chips is ultimately demand for electronics, and increasingly for electrified and instrumented everything. Three secular drivers underpin TI's markets. First, vehicle electrification and the rising electronic content per car: an EV carries far more power-management, sensing, and isolation silicon than an internal-combustion car, and even combustion cars are gaining electronic content every model year. Second, industrial automation and electrification: factory robotics, building systems, grid and renewable-energy infrastructure, and electrified equipment all add analog content. Third, the more recent and fast-growing data-center driver: AI servers need enormous, dense, efficient power delivery, and TI sells the power-management and signal-chain content around (not inside) the AI accelerators.

The analog semiconductor market is large and growing steadily rather than explosively - industry estimates put it in the mid-$80-billion range in 2026 rising toward roughly $95-100 billion over the following five years, a high-single-digit growth rate, with embedded processing adding further to TI's addressable market. TI sits near the centre of this supply chain as one of the few vertically integrated players that designs, manufactures, and packages its own chips, in contrast to fabless analog designers who depend on foundries.

The industry is cyclical, and the last few years are a textbook example. Analog rode a pandemic-era boom, then a sharp inventory correction through 2023-2024 as industrial and automotive customers worked down stockpiles, then a recovery that took hold through 2025 and into 2026 (industrial leading, automotive lagging). Across a full cycle, demand tracks industrial production and auto build rates plus the secular content-growth tailwind. The regulatory and policy layer matters more than it used to: the US CHIPS Act subsidises domestic fab construction (TI is a direct beneficiary), while tariffs and US-China trade tension cut both ways - they create reshoring/supply-security demand for US-made chips but also risk demand pull-ins, retaliation, and a faster build-out of a competing Chinese domestic analog industry. The tailwinds are content growth, electrification, data-center power, and reshoring; the headwinds are cyclicality, Chinese low-end competition, and trade-policy whiplash.

7. Growth triggers

All points below are drawn directly from the six concalls.

• Data center becoming a distinct, fast-growing market. Data center ran at roughly a $1.2 billion run rate in 2025 with growth exceeding 50% year to date, and management began breaking it out as a distinct end market (Q3 2025 concall, Oct 21 2025). By Q1 2026 data-center revenue was up about 90% year over year with sequential growth above 25% (Q1 2026 concall, Apr 22 2026).

  "The combination of broad portfolio, both on general purpose and ASSPs, ability to support the rack... is very, very unique." (Q1 2026 concall, Apr 22 2026)

• Application-specific data-center sockets ramping in H2 2026 and 2027. Management expects application-specific socket momentum to accelerate in the second half of 2026 and into 2027 (Q1 2026 concall, Apr 22 2026). New.
• Sherman 300mm fab in production. The first Sherman, Texas 300mm fab (SM1) began production in 2026, adding low-cost internal capacity (Q1 2026 concall, Apr 22 2026). Repeated build-out theme from prior calls.
• Elevated capex cycle completing by end of 2026. TI is in the final stretch of its 2020-2026 elevated capital cycle; 2026 capex guidance was cut to a $2-3 billion range (down from ~$5 billion in 2025), which mechanically lifts free cash flow as spending normalises (Q4 2025 concall, Jan 27 2026; Q1 2026 concall, Apr 22 2026).
• Free cash flow per share inflection toward ~$8 for 2026. With capex falling, management framed roughly $8 of free cash flow per share for 2026 as achievable on mid-to-high single-digit revenue growth (Q1 2026 concall, Apr 22 2026).

  "Very likely, we will... easily be that $8 free cash per share for 2026." (Haviv Ilan, Q1 2026 concall, Apr 22 2026)

• CHIPS Act direct funding inflows. TI is drawing a direct-funding agreement worth up to $1.6 billion; $555 million arrived in Q1 2026, $630 million cumulatively, with remaining tranches tied to milestones - a multi-quarter cash tailwind (Q1 2026 concall, Apr 22 2026; Q3 2025 concall, Oct 21 2025).
• Industrial recovery broadening. Industrial grew about 25% year over year in Q3 2025 and over 30% across all sectors and regions in Q1 2026, the clearest sign the cyclical recovery has taken hold (Q3 2025 concall, Oct 21 2025; Q1 2026 concall, Apr 22 2026). Repeated and strengthening.
• Automotive recovery still ahead (potential trigger). Automotive has lagged; management repeatedly flagged it as the next leg if it turns (Q2 2025, Jul 22 2025; Q1 2026, Apr 22 2026).

  "I want to see automotive... develop in Q2. It's too soon to call it." (Haviv Ilan, Q1 2026 concall, Apr 22 2026)

• Restructuring and 150mm/200mm fab wind-down benefits flowing through 2026. TI is winding down older 150mm and 250mm fabs; the cost benefits emerge gradually through 2026 (Q3 2025 concall, Oct 21 2025).

8. Key risks

Automotive remaining stuck in the doldrums. Industrial recovered through 2025-2026, but automotive stayed flat-to-shallow. Automotive is roughly 30% of revenue; if EV demand softness, inventory overhang, and a slow Western auto cycle persist, a third of the business does not participate in the recovery that the stock is pricing. Management has been candid that it cannot yet call the automotive turn (Q2 2025 concall, Jul 22 2025; Q1 2026 concall, Apr 22 2026), which is honest but also an admission that a large revenue pillar is not yet firing.

The capex bet could overshoot demand. TI committed tens of billions to owned 300mm capacity on the thesis that low-cost internal capacity wins over a cycle. If end-demand growth disappoints, TI will carry expensive, under-utilised fabs whose depreciation drags gross margin for years. The mechanism is simple: fixed manufacturing cost spread over fewer wafers raises unit cost and compresses margin. Management has hedged by trimming 2026 capex toward the low end of its framework (Q4 2025 concall, Jan 27 2026), which signals it is sizing the back end of the build to actual demand rather than building blind, but the capacity already committed is large relative to current revenue.

Chinese domestic competition in commodity analog. A well-funded, subsidised Chinese analog industry is climbing from the low end up. In the near term it commoditises pricing on basic power-management parts inside China; over time it threatens to displace foreign suppliers in the world's largest electronics market. Management has flagged China revenue volatility and the risk of demand being a pull-in rather than structural (Q2 2025 concall, Jul 22 2025). This is a slow, high-probability margin and share drag at the commodity end rather than a sudden catastrophe.

Tariff and trade-policy whiplash. TI sells globally and manufactures heavily in the US. Tariffs and retaliation can both pull demand forward (customers stockpiling ahead of tariffs, inflating a quarter that then gives back) and disrupt the China market. Management repeatedly cited tariffs as making customer ordering hard to read and warned that some quarters' strength may reflect temporary inventory loading (Q1 2025 concall, Apr 23 2025; Q2 2025 concall, Jul 22 2025). The risk is less existential than analytical: it makes near-term demand signals noisy and raises the odds of a disappointing quarter after a pull-in.

Cyclicality and inventory whips. Analog is cyclical, and TI carries high inventory by design to assure supply. A demand air-pocket would leave it correcting inventory and under-absorbing fabs simultaneously - the same dynamic that hurt it in 2023-2024. This is a recurring, moderate-severity feature of the business rather than a one-off.

9. Walk the talk

The six calls span the bottom and recovery of an analog cycle, which makes them a good test of management credibility - it is easy to look consistent in a boom and easy to look erratic in a bust.

At Q4 2024 (Jan 23 2025), with revenue down year over year and the cycle near its trough, management's central message was discipline-through-the-downturn: keep investing in 300mm, keep the 2020-2026 capex framework intact, and let the manufacturing advantage compound. They guided Q1 2025 revenue to $3.74-$4.06 billion and reiterated they were "nearly 70% through" the elevated capex cycle. Crucially they did not promise a sharp recovery; they promised to be ready for one.

By Q1 2025 (Apr 23 2025) the company delivered revenue around $4.1 billion, above the guided range, and guided Q2 up about 7% - better than normal seasonality. Management was careful, when pressed on tariffs, not to over-claim: CEO Haviv Ilan attributed the strength to shipping more volume rather than tariff-driven pull-ins and explicitly said he did not yet see a near-term tariff impact. This is the first instance of a pattern that repeats: guide cautiously, beat, and refuse to declare victory.

At Q2 2025 (Jul 22 2025) revenue came in around $4.4 billion, up 16% year over year with every end market except automotive growing. Management could have spun this as a clean recovery. Instead they flagged that some China strength (up ~32% year over year) might be tariff-related pull-in and called the automotive recovery "shallow." That is the opposite of overpromising - they actively talked down the quality of a good quarter.

Q3 2025 (Oct 21 2025) delivered ~$4.7 billion, up 14% year over year, with industrial up ~25% and data center disclosed as a ~$1.2 billion run rate growing over 50%. Management raised the dividend 4% (the 22nd consecutive annual increase) and committed to breaking out data center as a distinct market - a promise that is specific and checkable. Guidance for Q4 was $4.22-$4.58 billion.

"[We will provide] a distinct data center market revenue breakout starting next quarter." (Q3 2025 concall, Oct 21 2025)

At Q4 2025 (Jan 27 2026) revenue was ~$4.4 billion and the company missed slightly on EPS - a rare stumble, and notably one management did not paper over. They guided Q1 2026 to $4.32-$4.68 billion and, importantly, signalled 2026 capex would trend toward the low end of the framework ($2-3 billion versus ~$5 billion in 2025). This is the company adjusting the back end of its multi-year bet to demand rather than building stubbornly - consistent with the discipline they preached at the trough.

Then Q1 2026 (Apr 22 2026) was a clear beat: revenue ~$4.8 billion (above the top of the range), EPS $1.71 against guidance of $1.22-$1.48, Analog up 22%, data center up ~90% year over year and now formally broken out as promised two quarters earlier. Management guided Q2 to $5.0-$5.4 billion. The data-center breakout commitment from Q3 2025 was delivered on schedule.

The pattern across all six is consistent and favourable: management guides conservatively, generally beats, refuses to over-claim even on good quarters (the China pull-in and shallow-automotive caveats are the tell), and adjusts the capex plan to reality rather than to ego. They kept the specific, datable promise to break out data center. They were honest about the one EPS miss. The one open commitment is the automotive recovery, which they have repeatedly declined to call - which is itself credibility-preserving, because calling it early and being wrong would cost more than waiting. This is management that under-promises and delivers, not management that talks a big game.

10. Shareholder friendliness index

Dividends. TI is a serial dividend grower with one of the longest streaks in tech - 22 consecutive annual increases through 2025. The dividend rose from roughly $4.96 per share paid across 2023 (quarterly $1.24), to about $5.20 across 2024 (quarterly $1.30), to roughly $5.56 across 2025 (quarterly $1.36 stepping to $1.42 after the September 2025 4% raise) (TI investor relations dividend history; Q3 2025 concall, Oct 21 2025). The trajectory is steady mid-single-digit annual growth with no cuts. The one nuance worth noting: because TI is in the tail of a heavy capex cycle, free cash flow has been temporarily depressed and the dividend has at times consumed a high share of it - a function of the investment phase, not of distress, and a metric set to ease sharply as capex falls toward the guided ~$8 of free cash flow per share for 2026 (Q1 2026 concall, Apr 22 2026).

Buybacks and dilution. TI has historically been an aggressive repurchaser - it has reduced its share count materially over the past two decades by buying back stock alongside dividends. The MoatMap database recorded no buyback transactions in the trailing ~90 days (since 2026-03-22); this reflects a deliberate, disclosed pullback in repurchase pace during the elevated-capex/recovery phase, not the absence of a program. The fuller multi-year picture from filings: TI repurchased about $1.5 billion of stock in 2025, completed its longstanding 2018 share-repurchase authorization (~49.9 million shares for ~$7.7 billion in total under that authorization), and in Q1 2026 bought back only about $158 million - the company prioritised dividends and capex over buybacks while spending was elevated, returning roughly $6 billion total to shareholders over the trailing twelve months as of Q1 2026 (Q1 2026 concall, Apr 22 2026; TI 2025 capital-management disclosures). Net share count has been roughly flat-to-slightly-down recently as reduced buybacks offset modest option dilution, after years of steadier shrinkage.

Verdict: Returns Capital - a disciplined, dividend-first capital returner with a 22-year raise streak and a history of heavy buybacks, currently throttling repurchases through the back end of its capex cycle with the explicit aim of resuming heavier returns as free cash flow per share inflects upward.

11. Insider activities

Source: MoatMap database (US, SEC Form 4 spine), data current as of 2026-06-19, cross-checked against the disclosed pattern. The window covers the last 12 months. Across 26 transactions by 9 insiders, the tally is 0 open-market buys, 15 sells, and 11 "other" (option exercises, grants, deemed-interest/cost-basis legs). Net direction: selling, and the "other" rows are overwhelmingly option/RSU exercises paired with same-day sales - the low strike prices ($79-$175) against sale prices near $280-$321 confirm these are exercise-and-sell events, not conviction trades.

Buys - the signal: there were no open-market purchases by any insider in the window. There is no cluster buying, no first-time CEO or CFO purchase, nothing to flag as a bullish conviction signal. This is a neutral-to-mildly-cautious read on its own, though for a mega-cap where executives are compensated heavily in equity, the absence of open-market buying is unremarkable.

Sells - the why: every sell in the window fits the routine pattern of executives monetising vested equity compensation. The dead giveaway is structure: most rows pair an "other" leg (an option/RSU exercise at a low historical strike) with a same-day "sell" leg at the prevailing ~$280-$320 market price. These are equity-comp realisations, the kind typically executed under pre-arranged 10b5-1 plans, not directional bets against the business. The clustering in late April and mid-May coincides with the post-earnings trading window that opened after the Q1 2026 results (Apr 22 2026) - a standard window in which executives' scheduled sales execute. The CFO's (Lizardi) and CEO's (Ilan) sales are the largest by value but both are exercise-and-sell of long-dated options, not liquidation of core holdings. No estate, charitable, or sponsor-exit driver is indicated, and none is disclosed in the rows.

Net assessment: insiders are net sellers, but the selling is broad-based, routine equity-compensation monetisation concentrated in the post-earnings window, not concentrated emergency selling by one or two people. There is no open-market buying to provide a positive conviction signal, and no unusually aggressive or out-of-pattern selling to raise a red flag. Read: neutral. The insider tape here carries little informational content beyond the ordinary mechanics of executive pay.

12. Scenarios

Bull case. The cyclical recovery that began in industrial in 2025 broadens decisively into automotive through 2026 and 2027, so that the two-thirds of the business that is industrial-plus-automotive fires together for the first time since the boom. Data center, growing near 90% year over year and now a distinct market, compounds as TI's application-specific power and signal-chain sockets win designs around the next generations of AI accelerators, and the H2 2026 socket ramp management flagged materialises. Meanwhile the capex cycle ends on schedule: the Sherman fabs ramp, the 300mm internal mix climbs toward the high-90s percent of wafers, and the structural cost advantage drops to the bottom line just as revenue accelerates. Free cash flow per share inflects past $8 and keeps climbing as depreciation peaks and rolls over, and TI resumes heavy buybacks on top of its 23rd, 24th consecutive dividend raises. The US-manufacturing footprint becomes a durable selling point as customers pay up for supply security, and Chinese low-end competition stays contained to commodity parts TI is happy to cede. In this world TI exits the decade as the structurally lowest-cost, broadest-catalogue analog supplier, throwing off cash and returning nearly all of it.

Base case. Management delivers roughly what it has guided. Industrial recovery continues at a healthy clip, automotive turns gradually rather than sharply, and data center grows fast off a still-modest base. Capex falls to the $2-3 billion range in 2026 and free cash flow per share reaches around $8, lifting returns to shareholders without a dramatic buyback resumption while the company watches demand. The 300mm transition proceeds on plan, slowly widening the margin advantage. China remains a volatile, competitive market where TI holds the higher-value sockets and concedes some commodity share. Revenue grows mid-to-high single digits, the dividend streak extends, and the stock's story becomes "the capex cycle is paying off and free cash flow is normalising" rather than any single dramatic catalyst. Nothing breaks; nothing dramatically surprises. TI is a steadily compounding, cash-returning analog franchise emerging from a heavy investment decade.

Bear case. Automotive stays stuck - EV demand disappoints, Western auto build rates stay soft, and a third of revenue refuses to participate in the recovery. The industrial bounce turns out to have been partly tariff-driven pull-in and gives back, leaving a demand air-pocket just as the new Sherman capacity comes online. TI then faces the worst combination: under-utilised, newly built 300mm fabs whose depreciation drags gross margin for years, plus an inventory correction echoing 2023-2024. Chinese domestic analog suppliers, subsidised and improving, climb out of the commodity tier and start taking mainstream power-management share inside China, capping TI's largest growth market and pressuring prices industry-wide. Trade-policy whiplash makes each quarter hard to forecast and occasionally ugly. Free cash flow per share stalls below the $8 ambition, buybacks stay throttled, and the multi-billion-dollar capex bet looks, for a stretch, like capacity built for a demand curve that did not show up on time. The dividend is never at risk, but the equity story stalls into "expensive capacity waiting for a cycle."

13. Further reading

• Texas Instruments: Leading the Analog Chips Industry - MBI Deep Dives, 2024-04-22 [paid]