Analog and Mixed-Signal Semiconductors: A Sector Briefing (The Cestrian Circle Newsletter)

DISCLAIMER: This note is intended for US recipients only and, in particular, is not directed at, nor intended to be relied upon by any UK recipients. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Read our full disclaimer, here.

Analysis by Alex King, CEO, Cestrian Capital Research, Inc. + Claude CoWork

Introduction

This note examines the analog and mixed-signal semiconductor sector as a potential investment theme. The sector is large, technically distinct from digital logic and memory, and has a different demand profile from the GPU-centric AI narrative that has dominated semiconductor investor attention since 2023. Our assessment is that it remains largely undiscovered as a cohesive investment theme on financial social media and, to a meaningful extent, in institutional portfolios outside specialist technology funds.

I. What Analog and Mixed-Signal Semiconductors Are

The semiconductor industry divides broadly into digital and analog. Digital chips process information as discrete binary states — ones and zeros. Analog chips process continuous signals: voltage, current, temperature, pressure, light intensity, sound, radio frequency. The physical world operates in continuous variables. Every interface between that physical world and a digital computing system requires analog circuitry.

Mixed-signal devices combine analog and digital functions on a single die. The most common example is the analog-to-digital converter (ADC), which samples a continuous analog input and quantifies it as a stream of digital values. The reverse — digital-to-analog conversion (DAC) — is equally pervasive. A smartphone audio codec, a motor driver in an industrial robot, a battery management system in an electric vehicle, and a power regulator feeding a data center GPU are all mixed-signal devices in the broadest sense.

The principal product categories within analog and mixed-signal semiconductors are:

• Power management ICs — voltage regulators, DC-DC converters, battery management systems, gate drivers for power transistors. This is the largest single sub-category by revenue.
• Signal chain / data converters — ADCs, DACs, amplifiers, filters, comparators. Used wherever real-world signals are measured, processed, or generated.
• Interface ICs — USB, CAN, RS-485, I2C controllers, level translators. Enable communication between subsystems operating at different voltage levels or protocols.
• RF and microwave — amplifiers, mixers, filters, and transceivers operating at radio frequencies. Critical for cellular infrastructure, radar, satellite, and wireless connectivity.
• Motor drivers and motion control — integrated circuits that translate digital commands into precisely controlled current flows through electric motors.
• Sensor interface — front-end circuits that condition and digitize inputs from temperature sensors, accelerometers, pressure transducers, and similar devices.

The defining commercial characteristic of analog chips is long product life cycles combined with high switching costs. A design win — the process by which a customer qualifies an analog chip for use in a specific product — typically takes two to four years and, once secured, may generate revenue for a decade or more without redesign. This contrasts sharply with digital chips, where process node transitions drive frequent design refreshes. The consequence is high gross margins, sticky customer relationships, and relatively predictable long-term revenue streams — offset by slow revenue ramp and exposure to multi-year inventory cycles.

II. Market Size and Structure

The analog semiconductor market, broadly defined, may be valued at somewhere between $90–110 billion in 2025 depending on scope and methodology. The range reflects definitional differences: the narrower figure excludes power discretes and certain mixed-signal ASICs; the broader figure approaches the totality of chips whose primary function is interfacing with the physical world rather than executing digital computation. The compound annual growth rate across most published forecasts is 4.5–6% through 2033, implying a market of $150–175 billion by that date.

By end market, automotive and industrial each account for roughly 30% of demand, with consumer electronics, communications infrastructure, healthcare, and aerospace/defense dividing the remainder. The automotive share has risen materially over the past decade as vehicle electrification and advanced driver-assistance systems (ADAS) have driven analog content per vehicle from approximately $300 in a conventional internal combustion vehicle to $600–900 in a battery electric vehicle. Industrial automation and robotics represent the second major structural driver.

The sector has historically been concentrated among a small number of scaled players. Texas Instruments, Analog Devices, NXP Semiconductors, STMicroelectronics, Infineon, and Renesas collectively account for the majority of global analog revenue. Below them sits a tier of mid-cap specialists — Microchip Technology, ON Semiconductor, Monolithic Power Systems, MACOM Technology, and others — with more concentrated exposure to specific end markets or product categories.

III. The AI, Robotics, and Automation Thesis

The dominant semiconductor investment narrative since 2023 has been GPU-centric: Nvidia, HBM memory suppliers (SK Hynix, Micron), and optical interconnect companies have attracted the bulk of thematic capital and analyst attention. The analog connection to AI is real but indirect, and it is this indirectness that has kept the sector out of the spotlight.

Power Delivery at Scale

Every AI training cluster and inference server requires power management at multiple levels: from the AC supply through voltage regulators down to the millivolt-level power rails feeding individual GPU or ASIC dies. A high-end Nvidia H100 server rack contains thousands of analog power management devices. As data center power consumption scales — from gigawatts today toward estimates of 50+ gigawatts globally by the end of the decade — the demand for power management ICs scales with it, with specifications becoming progressively more demanding in terms of efficiency, current density, and response speed. Monolithic Power Systems is the most directly levered US-listed company to this theme within the analog space, given its strong position in AI server power delivery.

Edge AI and Sensor Fusion

The longer-term AI buildout moves computing to the edge: into vehicles, industrial equipment, medical devices, and consumer electronics. Edge AI systems cannot rely on cloud latency for real-time decision-making. Every edge AI node requires analog front-ends to process sensor data, power management to operate efficiently on constrained energy budgets, and signal chain components to interface with the physical environment. Texas Instruments, Analog Devices, and NXP are the principal beneficiaries of this trend at scale.

Electrification and Motor Control

Electric vehicles, industrial robots, and automation equipment all require precise control of electric motors. Motor control demands high-performance gate drivers, current sensing, and position feedback circuits — all analog or mixed-signal functions. The transition from hydraulic and pneumatic actuation to electric drives in industrial equipment represents a structural expansion of analog content per installed unit. onsemi, with its silicon carbide power device portfolio, is the most directly levered US-listed name to EV powertrain electrification within this space, though the EV demand slowdown of 2024–2025 has weighed heavily on its financials.

Robotics

A humanoid or industrial robot contains substantially more analog content than a conventional machine: inertial measurement units, torque sensors, proximity sensing, motor feedback, battery management, and audio/visual processing front-ends. The autonomous systems and robotics segment is projected to grow at the fastest rate within analog end markets through 2030. While robot volumes remain small relative to automotive and industrial in aggregate, the analog content per unit is high, and the design cycle dynamics favor suppliers who can offer integrated, application-specific solutions.

The Indirect Nature of the Exposure

It is important to be precise: analog chips do not perform AI computation. The silicon doing the arithmetic is digital — GPUs, TPUs, custom ASICs. Analog chips enable those digital systems: they convert sensor inputs, deliver clean power, control motors, and handle wireless connectivity. The exposure is therefore best characterized as infrastructure-level rather than application-level. This makes it less exciting as a narrative — analog does not generate the revenue multiples of an Nvidia — but it does make it more durable. Analog demand is broadly proportional to units of intelligent, electrified, and connected equipment deployed in the world, rather than to the compute intensity of specific workloads.

IV. The Inventory Cycle Context

Any assessment of the sector must be read against the backdrop of the most severe analog inventory correction in recent memory. Following the demand surge and supply-chain panic of 2021–2022, customers across automotive and industrial markets built excess component inventory. The destocking cycle ran through 2023 and 2024, depressing analog revenues significantly across the sector:

• Texas Instruments: FY2024 revenue of $15.6 billion, down 11% year-on-year
• Analog Devices: FY2024 revenue of $9.4 billion, down 23% year-on-year
• Microchip Technology: FY2024 revenue of $7.6 billion, down 10% year-on-year
• ON Semiconductor: FY2024 revenue of $7.1 billion, down 14% year-on-year
• NXP Semiconductors: FY2025 revenue of approximately $12.3 billion, declining from 2024 levels

Recovery is now underway but uneven. Texas Instruments returned to growth in 2025 with revenue of $17.7 billion, up 13%. Analog Devices reported FY2025 revenue of $11.0 billion, up 17%, with free cash flow of $4.3 billion on 39% of revenue — among the highest free cash flow margins in the semiconductor industry. Companies with heavier EV exposure, notably onsemi, continued to struggle: FY2025 revenue of $6.0 billion represented a further 15% decline, as EV production growth slowed materially from prior expectations.

The inventory cycle matters for investors because it has compressed trailing multiples and depressed analyst estimates relative to what recovery-cycle revenues will likely support. TXN and ADI are recovering faster; MCHP and ON are not yet through the trough. This creates a heterogeneous opportunity set within a sector that is often discussed as monolithic.

V. Principal US-Listed Companies and ADRs

The following covers the main investable names.