Texas Instruments

Texas Instruments designs, manufactures and sells analog and embedded semiconductors used to manage power, handle physical signals and control electronic equipment. Its business combines a broad catalog, internal manufacturing and direct customer relationships.

Texas Instruments Incorporated ·Nasdaq: TXN

30-Second Explanation

Texas Instruments is a semiconductor company focused on Analog and Embedded Processing rather than leading-edge CPUs or GPUs.

FY2025 revenue was $17.682B: Analog $14.01B, Embedded Processing $2.70B, and Other about $0.98B. Analog and Embedded generated roughly 95% of revenue. [S1]

TI sells 80,000+ products to 100,000+ customers. [S8] The business is intentionally diversified across thousands of applications.

The model is:

broad long-lived portfolio + direct customer relationships + internal process technology + 300mm manufacturing + inventory availability ↓ many design wins across many customers ↓ high gross margin and cash generation ↓ R&D + fabs + shareholder returns

Latest Financial State — Q3 2026

Q3 2026 revenue was $6.055B, operating profit $2.874B, net income $2.425B and diluted EPS $2.62. [S2]

Segment results:

Analog              $4.911B revenue / $2.505B operating profit
Embedded              $858M revenue /   $227M operating profit
Other                 $286M revenue /   $142M operating profit

Analog grew 32% year over year and Embedded 21%. Management described broad growth across industrial, automotive and data center. [S46]

Q4 guidance was $5.8B–$6.3B revenue and $2.45–$2.79 EPS. [S45]

Why Analog Is Different

Analog semiconductors perform physical/electrical functions: voltage regulation, battery management, amplification, conversion, isolation, interface and sensing. Almost every electronic system needs multiple analog chips.

Unlike a flagship processor, an analog part often does not become obsolete because a newer transistor node appears. A qualified amplifier or power regulator can stay in a customer design for many years. This supports long product lives, long inventory usefulness and high cumulative returns on one design/R&D investment.

Embedded Processing Economics

Embedded products provide control and computation through MCUs, processors, connectivity and radar. Customer-written software creates switching cost. Once a factory controller or vehicle ECU has years of validated code, changing the MCU means more than swapping a chip: software, safety, testing and tools may need rework. [S11][S12]

2025 End-Market Mix

TI's 2025 market mix was approximately Industrial 33%, Automotive 33%, Data Center 9%, Personal Electronics 21% and Communications Equipment 3%. [S42]

Industrial + Automotive + Data Center represented about 75% of revenue and receive additional strategic emphasis because semiconductor content and product lifetimes can be attractive.

The Long-Tail Portfolio

The portfolio has more than 80,000 products. [S8][S9] A customer may use TI for an MCU, multiple regulators, an ADC, an amplifier, isolation and interface parts in one design. Breadth raises the probability that TI participates in more sockets without requiring any single chip to dominate company revenue.

This long tail reduces dependence on one end customer or one product cycle. The 2025 report says about half of revenue came from customers outside the largest 50. [S41]

Direct Customer Relationships

More than 80% of 2025 revenue was direct, versus roughly one-third in 2019. [S7][S41]

TI invested in TI.com, order fulfillment, inventory programs, logistics and direct sales. The goal was to see more design activity directly and sell more TI parts into each design. This gives TI better demand information and pricing/customer insight, but also requires TI to operate more channel infrastructure itself.

Internal Manufacturing

TI's manufacturing strategy is unusually integrated. It operates 15 manufacturing sites and targets more than 95% internally manufactured wafers by 2030, with more than 80% on 300mm. [S19][S40]

The advantage is lower structural cost and greater supply control. The cost is fixed assets: when demand is weak, depreciation and underutilization pressure margins.

300mm Footprint

Richardson: RFAB1 and RFAB2 are in production. RFAB1 pioneered 300mm analog manufacturing. [S21]

Lehi: LFAB1 is in production; LFAB2 is under construction. [S22]

Sherman: SM1 entered production in December 2025; the site can ultimately host four connected fabs and SM2 construction has advanced. [S20][S24]

TI's long-term U.S. manufacturing plan contemplates more than $60B of potential investment across Texas and Utah sites. [S23]

Why 300mm Can Lower Cost

A 300mm wafer has much more surface area than a 200mm wafer. If a process/product can migrate efficiently, more die are produced per wafer and automation/fixed factory cost is spread across more chips.

Analog economics are especially suited to this because products can run for years on defined process technologies. TI can invest once in a manufacturing platform and reuse it across thousands of products.

Capex Cycle and FCF Recovery

2025 cash flow from operations was $7.153B, capex $4.550B and free cash flow under TI's definition $2.938B after CHIPS proceeds. [S47]

For the trailing twelve months ending Q3 2026, CFO rose to $9.896B, capex fell to $2.578B and free cash flow rose to $8.422B after $1.104B of CHIPS incentive proceeds. [S4][S48]

The strategic sequence is:

build fabs ahead of demand ↓ high capex / low near-term FCF ↓ capacity becomes available ↓ demand + factory utilization recover ↓ capex normalizes ↓ FCF expands

Inventory Strategy

Year-end 2025 inventory was $4.80B and days inventory 222. [S39]

TI intentionally holds more inventory because many Analog/Embedded products have long lives. Inventory lets TI support fragmented customers quickly when demand rebounds. The cost is working capital and the risk that forecasts/product lifecycles are wrong.

Gross Margin and Factory Loading

Q3 2026 gross profit was $3.887B on $6.055B revenue, about 64.2% gross margin. [S2]

Factory loading matters because internal manufacturing has substantial fixed cost. In a downturn, lower utilization raises unit cost. In recovery, higher volume can produce strong operating leverage as the same factory depreciation/overhead is spread across more units.

Automotive

Automotive represented about 33% of 2025 revenue. [S42] Semiconductor content rises through EV power conversion, battery management, radar, ADAS, networking, body electronics and safety.

TI's 2026 portfolio included high-performance automotive SoCs, 8x8 4D imaging radar and 10BASE-T1S automotive Ethernet. [S14]

Industrial

Industrial represented about 33% of 2025 revenue. [S42] It spans automation, robotics, energy infrastructure, medical, aerospace/defense, buildings and test/measurement. Long equipment lives and fragmented customers fit TI's catalog, inventory and direct-channel strengths.

Data Center

Data Center represented about 9% of 2025 revenue after TI separated it from the old enterprise-systems category. [S42]

AI servers need power conversion, hot-swap protection, sensing, clocks, isolation and control around the GPU/CPU. TI can benefit from AI infrastructure without making the accelerator itself.

Capital Allocation

TI says long-term free cash flow per share is its ultimate value metric. It invests first in competitive advantages such as R&D and manufacturing, then returns free cash flow to owners over time.

TTM Q3 2026: dividends $5.174B, repurchases $0.636B, total cash returned $5.810B. [S44]

Silicon Labs Pending Acquisition

In February 2026 TI agreed to acquire Silicon Labs for $231/share, about $7.5B enterprise value. [S35] Silicon Labs adds about 1,200 wireless-connectivity products. TI targets roughly $450M of annual manufacturing/operational synergies within three years post-close.

At Sept. 30, 2026 the deal was still pending, expected to close in H1 2027 subject to approvals. [S36]

The strategic fit is direct:

Silicon Labs wireless IP + TI Embedded portfolio + TI direct customer channel + TI internal process and 300mm manufacturing ↓ more embedded sockets + lower manufacturing cost + cross-sell

Competition

Analog and embedded markets remain fragmented. Competitors include Analog Devices, Infineon, STMicroelectronics, NXP, Microchip, Renesas, onsemi, Monolithic Power Systems and many niche/Asian suppliers.

TI's defense is not a monopoly architecture. It is the combined breadth of product, manufacturing, channel and long design life.

History

1930: Geophysical Service Inc. founded. 1951: Texas Instruments name adopted. 1954: TI developed an early commercial silicon transistor. 1958: Jack Kilby demonstrated the first working integrated circuit. 1967: handheld calculator. 1971: single-chip MCU milestone. 1980: commercial single-chip DSP. 1985: DLP micromirror invention. [S8][S28]

2000: Burr-Brown acquisition. 2009: Luminary Micro. 2011: National Semiconductor. 2021 onward: major 300mm expansion. 2023: Haviv Ilan became CEO. 2026: Ilan became chairman and TI announced Silicon Labs. [S26][S30][S33][S34][S35]

Common Misconceptions

  • TI is not mainly a calculator company; calculators sit in Other.
  • Mature-node analog does not mean no growth; electronics need more power/sensing/interface content.
  • TI's fabs are primarily for TI's own chips, not a TSMC-like merchant foundry.
  • High inventory can be strategic when product lives are long.
  • Silicon Labs is not yet consolidated at the research cutoff.
  • 300mm lowers unit cost only when technology, yield and utilization support the economics.

Key 2026–2027 Watchlist

Cycle recovery

Does industrial/automotive/data-center growth remain broad after the sharp Q3 rebound?

Utilization

Can new 300mm capacity fill fast enough to improve fixed-cost absorption?

Capex

Does annual capex stay near the lower post-buildout range before the next equipment wave?

Silicon Labs

Does regulatory review stay on track for H1 2027?

Embedded

Can TI materially accelerate Embedded growth beyond its historical scale after the acquisition?

Competition

Do Chinese/regional analog/MCU suppliers pressure pricing in general-purpose products?

Shareholder returns

Can FCF support the dividend, repurchases and acquisition financing simultaneously?

How the Company Actually Works

TI begins with an equipment requirement: regulate power, measure a signal, control a process, move information or perform embedded computation. Its engineers turn those requirements into reusable semiconductor products supported by documentation, software where relevant and qualified manufacturing routes.

The customer evaluates the part inside a circuit or software platform. A successful evaluation may lead to qualification and a production design, but not necessarily immediate revenue. Equipment schedules and customer inventory determine when components are ordered.

customer electrical or computing requirement
↓
TI product, process and application engineering
↓
customer evaluation and qualification
↓
manufacturing, assembly and test
↓
inventory and order fulfillment
↓
component shipments for customer equipment

The broad catalog makes this sequence repeat across many applications. The economic opportunity is the sum of useful design positions rather than one device that controls the entire customer system. A production design can use TI and competing suppliers together.

Who Uses the Products and Who Pays

Design engineers choose electrical behavior, peripherals and software fit. Systems teams define the equipment's operating constraints. Quality and reliability teams influence qualification. Procurement or a manufacturing partner buys the approved parts.

The ultimate user can be a driver, factory operator, patient, technician or data-center customer who never sees TI's name. TI generally earns component revenue earlier in the chain rather than a share of the final equipment sale or service subscription.

This separation explains why demand is difficult to read. The purchaser can be several steps away from final use, and its inventory policy can move orders ahead of or behind equipment consumption. Direct customer relationships provide more information without removing that uncertainty.

Money Flow and the Cost of Serving Designs

customer or manufacturing partner pays for components ↓ TI recognizes product revenue under the applicable sales terms ↓ production, development and commercial costs are covered ↓ operating cash supports investment and other obligations

The flow is an explanation of the model, not a claim that cash and accounting revenue arrive simultaneously in every transaction. Receivables, inventory and payment timing can separate earnings from cash.

An attractive design can generate repeat shipments over years, but it is not a guaranteed subscription. Developing and maintaining the product, supplying samples, supporting software and preserving qualified routes all consume resources. The whole lifetime relationship matters more than one unit's apparent gross profit.

Dependencies and Bottlenecks

TI depends on process tools, materials, energy, logistics and skilled engineering even when manufacturing is internal. Customers depend on usable specifications, software and reliable component delivery. These are different sides of the same operating system.

The bottleneck can change. More wafer capacity does not solve an inadequate software platform or an assembly constraint. More website traffic does not prove enough qualified demand exists to load a fab. Matching resources across the chain is the central management task.

The 2026 capital-management materials link business-model advantages and capital decisions. Those materials help explain management's framework, but a stated objective is not evidence that every investment has already earned the intended return. [S5][S6]

Competition by Business Layer

LayerWhat customers compareWhy one ranking is insufficient
PowerEfficiency, protection, package and system costDifferent voltage and load requirements create different contests
Signal ChainUseful accuracy, noise, speed and integrationResolution alone does not establish measurement performance
MCU and controlPeripherals, timing, power and reusable softwareA common CPU core does not make devices interchangeable
Application computeWorkload performance, interfaces and toolsPeak arithmetic capability is not complete application throughput
Connectivity and radarProtocol or sensing fit, software and integrationCommunication and environmental sensing solve different problems
Supply and channelQualified availability, support and order executionCatalog breadth is valuable only when customers can use it

TI's combined portfolio can reduce search and integration work. Specialists can still win difficult sockets, and integrated alternatives can remove separate devices. Manufacturing scale and channel reach strengthen competitive products rather than replace technical fit.

Leadership and Organizational Continuity

Haviv Ilan became CEO in 2023 and chairman in January 2026. Rich Templeton retired as chairman at the end of 2025. The formal succession separates current operating responsibility from the earlier construction of the strategy. [S26]

The company needs more than one executive's judgment. Product teams, software developers, application engineers and manufacturing operations make different parts of the model work. A broad strategy becomes effective when those teams exchange useful customer and production information rather than optimize isolated metrics.

Common Questions

Why is TI still relevant when it does not make the main AI accelerator?

Computing equipment also needs electrical power, monitoring, interfaces and control. TI can participate through those functions. The opportunity is supporting component sales, not the full value of AI infrastructure.

Are end-market shares additional to segment sales?

No. Product segments, market applications and sales channels describe overlapping views of the same underlying revenue. Their totals must not be added together.

Does lower capex prove a permanent improvement in the business?

It improves the current cash calculation but does not by itself establish stronger customer demand, better products or sufficient future capacity. Operating cash, investment timing and incentive proceeds need separate interpretation.

Is the Silicon Labs transaction already complete?

Not at the October 31, 2026 research cutoff. The Q3 evidence describes a pending agreement and an expected future closing window. [S36]

Evidence

Connected Entities