Guide

The analog and power lane in India: what has been approved, what already runs, and what the record supports

Published 25 August 2026. Facts verified to 25 August 2026 unless dated otherwise.

Analog and power devices are what mature nodes were built to make, and India's approval book holds more of them than the headline numbers suggest.

The lane this page covers, and why it is not the logic race

Most coverage of India's semiconductor programme is written as though the industry had one product, the processor, and one measure of merit, the node number. Analog and power devices sit outside that frame entirely. Analog integrated circuits handle continuous quantities: they regulate voltage, sense current, condition signals, drive motors and manage batteries. Power devices switch and convert energy rather than compute with it, and the family covers MOSFETs, insulated gate bipolar transistors, rectifiers, thyristors and protection diodes. Neither category gets meaningfully better by shrinking the transistor, which is the single fact that explains almost everything else about how these plants are built, financed and judged.

This matters to a record that tracks India's build-out because the approved portfolio contains several projects in this lane, all of them small enough to be skipped in national coverage. The largest number in India's semiconductor book belongs to a silicon logic and mixed-signal fab. The smallest belongs to a power-device plant that has been operating for decades. Reading the second one as a rounding error is a mistake, because capital intensity in this industry varies by three orders of magnitude, and the cheapest projects are frequently the ones already shipping product.

Everything below carries a source and a date. Where the government has published a figure, this page uses the government's figure. Where a company has published its own claim about itself, the claim is marked as the company's and tiered accordingly. Where nothing has been published, this page says so instead of estimating.

Why mature nodes are the natural home of analog and power

Digital logic improves when transistors get smaller: more of them fit, they switch faster, and the cost per function falls. Analog does not follow that curve. An analog circuit depends on voltage headroom, on how precisely two devices on the same die match each other, on the quality of resistors and capacitors built alongside the transistors, and on the thickness of the metal that carries current. Shrinking a process lowers the supply voltage the transistors can survive, tightens the geometry in ways that make matching harder, and offers nothing an amplifier or a regulator wants. Porting an analog block down two nodes usually makes it worse, larger in real terms, and more expensive to qualify. Our definitional answer covers the vocabulary that goes with this.

Power devices are further still from the shrink argument. A device that must block hundreds of volts and carry tens of amperes is designed around drift regions, thermal paths and package resistance. Its performance is set by material properties and structure rather than by lithographic resolution, which is why power discretes are still made on 150mm and 200mm lines, and why the industry's power leaders run fabs that were paid for long ago.

The commercial consequence is the part worth internalising. Because these products do not chase the node curve, their fabs keep earning after depreciation ends, their product lifetimes run for a decade or more, and their customers qualify a part once and then buy it for years. Automotive and industrial buyers in particular treat a qualified part as an asset and change it reluctantly. A line that would be uncompetitive for a phone processor can be a durable business in analog and power, which is the honest structural argument behind India's entry point. Our mature node file covers those economics in full.

What India has actually approved in this lane

Four of the twelve approved projects touch analog or power directly. The figures below are the government's own, taken from the annexure to a Lok Sabha reply submitted by the Minister of State for Electronics and Information Technology on 1 April 2026 and published by the Press Information Bureau, from the Cabinet release of 12 August 2025 covering four units in Odisha, Punjab and Andhra Pradesh at an outlay of about Rs 4,600 crore, and from the Cabinet release of 5 May 2026 covering two further units in Gujarat.

Two readings follow. The first is that the analog and power lane in India is real but small in capital terms: two of the four projects were approved at under Rs 2,100 crore each, and the third has no individual figure in the public record at all, because the government published a combined Rs 3,936 crore for the pair cleared on 5 May 2026 without splitting it. The second is that the lane is disproportionately weighted toward assembly and discrete devices rather than analog integrated circuit fabrication, which is a different business with different equipment. Anyone reading the table as evidence that India is about to make precision analog integrated circuits at scale is reading past what the approvals actually say.

CDIL at Mohali: the plant that was already here

Continental Device India is the oldest entry in the portfolio by a wide margin. The company states on its own site, checked by this record on 25 August 2026, that it has operated since 1964, that its facilities are spread between Delhi and Mohali across four semiconductor manufacturing units, and that its product range covers transistors, diodes, rectifiers, thyristors, transient voltage suppressors and regulators in metal can and custom packages REPORTED. In September 2023 the company announced that it had become India's first manufacturer of silicon carbide components, alongside a new surface mount packaging line, and in February 2025 it announced indigenous solar bypass diodes REPORTED. Those are company statements about the company, which is exactly the tier they carry here.

The approval record adds the part the government vouches for. CDIL was cleared under the India Semiconductor Mission in August 2025 for a brownfield expansion of its Mohali discrete manufacturing at about Rs 117 crore, with an annual capacity of roughly 158.38 million units of high power discretes covering MOSFETs, IGBTs, Schottky bypass diodes and transistors in both silicon and silicon carbide DURABLE. The Press Information Bureau's own description of the applications for that batch of approvals names automotive electronics including electric vehicles and their charging infrastructure, renewable energy systems, power conversion, industrial applications and communication infrastructure.

The instructive number is the smallest one. Rs 117 crore is about one eight-hundredth of the capital approved for the Dholera fab, and it buys real additional output at an existing plant with existing people, existing utilities and existing customer qualifications. That is what a brownfield expansion in the discrete business costs, and it is the clearest single illustration of why capital intensity, not ambition, decides the order in which a country builds this industry. The capital ladder file works that comparison through properly.

The Dholera fab's analog content, stated carefully

The flagship project is not an analog fab, but its approved product set includes analog and power management work. The Union Cabinet cleared the Tata Electronics plant with Powerchip Semiconductor Manufacturing Corporation on 29 February 2024 at Rs 91,000 crore, the Fiscal Support Agreement of 5 March 2025 cites Rs 91,526 crore with central support covering 50 percent of eligible project cost, and the approved scope names power management chips, display drivers, microcontrollers and logic across a 28 to 110nm band at up to 50,000 wafer starts per month on 300mm wafers DURABLE. Power management integrated circuits are a canonical mature-node product, which is why they appear first in that list rather than as an afterthought.

The opening node is disputed on the record and this page does not resolve it. Bloomberg reported on 17 July 2026 that the plant will open mostly at 90nm, TrendForce carried the same reading on 20 July 2026, and the company's stated node range remains 28 to 110nm, while the Tata Sons chairman's letter in the FY25 annual report said the fab would start at 28nm REPORTED. For an analog and power reader the contradiction lands differently than it does for a logic reader. A 90nm opening would be unremarkable for power management parts, since that class of product is routinely built on nodes older than 90nm, and it would matter mainly against expectations set for finer-geometry work. Our node investigation holds the full documentary trail, and the company file on this desk holds the corporate record.

On timing, no first-silicon date is supported by any source this record has seen. Commercial production is guided to mid-2028 by the electronics and information technology minister, Ashwini Vaishnaw, on 17 July 2026 TARGET. Civil work stood at about 50 percent as of mid-2026 with cleanroom fit-out under way.

What India has not approved, stated plainly

There is no dedicated merchant analog foundry in the approved portfolio. The government's own categorisation, published when the Union Cabinet approved Semicon 2.0 on 15 July 2026 with a budget outlay of Rs 1,27,500 crore, describes the twelve approved manufacturing units carrying more than Rs 1.64 lakh crore as one silicon fab, one silicon carbide fab, one integrated gallium nitride micro light emitting diode display fab and nine packaging units DURABLE. Exactly one silicon wafer fab sits in that book, and it is the Dholera plant. Any claim that India has approved a separate analog fab does not survive the government's own arithmetic.

The other operating fabrication line in the country is the Semi-Conductor Laboratory at Mohali, which is a government facility rather than an India Semiconductor Mission project. Its working node is documented in the same Lok Sabha reply of 1 April 2026, which records that of 211 chips taped out by 75 Indian academic institutions using state-provided design tools, 149 were fabricated at 180nm at the Mohali laboratory and 62 went to overseas foundries. A 180nm line is a perfectly reasonable analog and power node in engineering terms, and the constraint at Mohali has been institutional capacity rather than node capability. Our Mohali file carries that record in full, and the fabs guide sorts fabs from packaging plants across the whole portfolio.

Three further gaps are worth naming because they are frequently assumed away. This record has seen no public announcement of a bipolar-CMOS-DMOS process line in India, which is the workhorse platform for automotive power management worldwide. No privately owned Indian fab offers analog process technology as a merchant service today, and the one shared-access route that exists runs through the government laboratory at Mohali, where the ChipIN Centre organised six shared wafer runs enabling 122 chip design submissions from 46 institutions. And no approved project in this lane has published a qualified automotive customer. The absence of an announcement is not proof that nothing is happening, but a record that prints only what is sourced has to leave the space empty until something fills it.

The demand argument, and where it stops being evidence

The demand case for analog and power in India is the strongest part of this lane, and it is also the part most often stretched. Electric vehicles need traction inverters, on-board chargers and battery management, all of which are power-device intensive. Charging infrastructure needs rectification and conversion. Solar installations need inverters and bypass diodes. Motor drives, industrial automation, rail electrification, appliances and telecom power supplies all consume the same families of parts. India's manufacturing and energy build-out therefore creates genuine domestic consumption of exactly the devices these plants make, and the Press Information Bureau's own application list for the August 2025 approvals says as much.

The limit is that national demand is not the same as orders for a specific plant. Power and analog parts sell into qualified supply chains, and qualification is slow by design: automotive programmes in particular run multi-year validation cycles, and buyers commonly insist on a second source before designing a part in. A plant that starts production in 2027 or 2028 is bidding for sockets in vehicles that are being designed now, which means the commercial work has to happen years ahead of the first shipment and is almost never disclosed while it is happening. Import substitution arithmetic that multiplies India's device consumption by a plant's nameplate capacity ignores every one of those gates.

What this record will watch

• Fiscal support agreements. Cabinet approval is not disbursement. The Dholera fab took about twelve months from approval in February 2024 to its agreement of 5 March 2025, which is the benchmark this record uses when reading the 2025 and 2026 approvals in this lane.
• Construction evidence at Surat. Suchi Semicon's discrete assembly plant was approved on 5 May 2026 with a proposed 1,033.20 million chips a year for power electronics, analog integrated circuits and industrial systems DURABLE. Physical progress during 2026 and 2027 is the test.
• Whether CDIL's expanded capacity appears. A brownfield line reaching its approved 158.38 million units a year would be the fastest measurable output gain in the portfolio.
• Semicon 2.0 instruments. The second phase covers design, manufacturing, research, materials, packaging and talent. Whether any of its materials or equipment support reaches power packaging, substrates or wide-bandgap materials will tell this record whether the analog and power lane is being treated as a policy priority or as a by-product.
• A first named customer. No approved plant in this lane has published a qualified customer. The first credible customer announcement, with a named counterparty and a product, would be worth more than any additional capacity headline.

The honest summary is that India has approved a modest, coherent set of analog and power projects, that one of them is an existing plant expanding rather than a greenfield promise, and that the country's single silicon fab will carry power management work inside a much broader portfolio. That is a smaller story than the headlines and a more durable one, because the products involved outlive the news cycles that announce them. The neighbouring lane, compound semiconductors, is covered in the compound file, which handles silicon carbide and gallium nitride as materials rather than as market slogans.