Company file

Fujifilm India

Published 3 August 2026. Facts verified to 27 July 2026 unless dated otherwise.

Fujifilm India signed an MoU with the Gujarat State Electronics Mission on 30 June 2026 to explore a semiconductor materials production base at Dholera.

Status tier: REPORTED

Read the verb, not the headline

Some coverage of this MoU rounds "explore" up to "will build". This record will not. An exploratory MoU is a company spending real diligence effort on a site, which is meaningful, and it is also still a company that can walk away, which is why the tier stays REPORTED with the word exploratory attached.

Why it still matters

Photoresists, CMP slurries and specialty films are the quiet half of a semiconductor ecosystem. Materials makers site their plants near confirmed fabs, so a global materials name running the numbers on Dholera is a direct consequence of the Tata anchor being real. If it converts, the ledger upgrades it. If it lapses, the ledger says so. Either way the promise vs delivery page keeps the score.

What a semiconductor materials supplier actually sells

It helps to separate the three things a fab buys. It buys tools, which are capital equipment bought once and written down over years of production. It buys facilities, which are built once. And it buys consumables, which it purchases every single day the line is running. Materials sit in that third category, and the distinction matters commercially, because a materials supplier's revenue tracks wafer starts rather than capital expenditure. A tool vendor is paid when a fab is built. A materials vendor is paid when a fab is busy.

The consumable list is longer than most people assume. Photoresist is the headline product: a polymer resin, a photoactive compound and a solvent system, spun onto the wafer in a film a fraction of a micron thick, which changes solubility where light strikes it. Different wavelengths need chemically different resists. Older i-line layers patterned at 365nm typically use novolak resins with diazonaphthoquinone sensitisers, while deep ultraviolet layers at 248nm and 193nm use chemically amplified formulations built around photoacid generators, where a single absorbed photon triggers a catalytic cascade. Around the resist sit anti-reflective coatings, developers, edge bead removers, strippers and rinse chemistries, all of which have to be compatible with one another and with the underlying film stack.

Then there is chemical mechanical planarisation. CMP slurries are engineered suspensions, commonly colloidal silica or ceria abrasives carried in a chemistry containing oxidisers, chelating agents, surfactants and corrosion inhibitors. The chemistry softens or converts the surface and the abrasive removes it, and the ratio between the two determines selectivity between materials, dishing, erosion and defectivity. Beyond that sit wet etchants, high purity solvents and acids, deposition precursors for CVD and ALD, photomasks and the blanks that mask shops build them from, and a family of specialty films: dicing tape, back-grinding tape, protective and surface-protection films used through assembly and test. Fujifilm's electronic materials business spans much of that consumables layer, which is the relevant context for a materials base rather than an equipment one.

Why the materials layer is one of the most concentrated in the industry

Very few industrial supply chains are as narrow as this one. A handful of mostly Japanese, American, German and Korean firms account for the bulk of global supply in several of these product categories, and the reasons are structural rather than accidental.

The first reason is purity. Metallic contamination in a resist or a wet chemical is specified in parts per billion and, for some species, parts per trillion, with particle counts controlled per millilitre above a given size. Achieving that is not a matter of buying better filters. It touches reactor metallurgy, solvent sourcing, packaging, drum liners, transfer lines and the training of the people who handle it. A trace metal that would be invisible in any other chemical industry becomes a yield-killing defect here.

The second reason is qualification. A fab does not simply switch suppliers. Introducing a new resist or slurry means running split lots, characterising the process window, monitoring defectivity, checking electrical parametrics and then sitting through reliability testing. That cycle is measured in many months and sometimes longer, and it consumes engineering capacity the fab would rather spend on yield. The consequence cuts both ways. An incumbent supplier is extremely sticky once designed in. A new entrant has almost no route in, because the customer bears the risk of trying it.

The third reason is that the addressable market for any single molecule is small relative to the development cost. A formulation is co-developed with a specific process on specific tools, and much of the value sits in tacit knowledge accumulated over decades of failure analysis rather than in a patent that can be read. That combination, high qualification friction and modest revenue per product line, deters entry more effectively than any barrier a government can build or remove.

Why materials firms follow fabs to the site

Bulk gases are generally made next door, which is why INOX Air Products began construction in October 2025 of a Rs 500 crore electronic specialty gas hub at Dholera, including a 200 tonne per day air separation unit DURABLE. Nitrogen at fab volumes is normally generated on site rather than trucked in. Materials sit in an awkward middle position: they can be shipped internationally, and for most of the industry's history they have been, but shipping them carries a set of costs that grow with volume.

Photoresists and many formulated chemistries are light sensitive and temperature sensitive, with usable storage life measured in months rather than years, which means refrigerated logistics and disciplined stock rotation. Flammable solvents and corrosives are dangerous goods, which restricts air freight and complicates every border crossing. Customs variability is a particular problem because a fab runs a just-in-time consumption model and does not want a large hazardous inventory sitting on site. Bulk chemical distribution systems inside a fab are designed around predictable replenishment, and returnable drums and totes create a reverse logistics loop that is much cheaper over a short distance.

There is also a service argument that is easily overlooked. When defectivity moves, the materials supplier's applications engineers are expected in the fab, looking at the same wafers as the process team. That relationship is difficult to run across a time zone and a visa regime. Proximity converts a vendor into a co-development partner. The trigger for building locally, however, is normally a confirmed volume anchor. Materials plants rarely create demand, they follow it.

How to read the verb "explore"

The Fujifilm memorandum with the Gujarat State Electronics Mission, signed on 30 June 2026, is described as an agreement to explore a semiconductor materials base REPORTED. That verb is doing a great deal of work and should be read literally rather than generously.

There is a ladder that runs from announcement to production, and each rung carries a different level of commitment. A memorandum of understanding sits at the bottom. It typically commits no capital, names no site, and is often explicitly non-binding, as the Intel memorandum of December 2025 was stated to be REPORTED. Above that sits a feasibility study, then land allotment or a lease, then an internal board sanction for capital, then environmental clearance and consent to establish, then an EPC award, then construction, then commissioning, and only then qualification with a customer. On the Dholera fab itself, the equivalent of a firm rung was the fiscal support agreement signed on 5 March 2025 DURABLE, which is a materially different instrument from a memorandum.

The counterparty matters too. A state electronics mission is a facilitation and promotion body. It can offer land, clearances and incentives. It cannot offer purchase orders. The customer for a materials plant is the fab, and no supply relationship between Fujifilm and any Dholera fab operator has been made public. An exploratory memorandum with a state agency is best understood as a low-cost option: it costs the company almost nothing, it gives the state a pipeline entry, and it keeps a seat at the table if volumes materialise.

What ISM 2.0's materials focus does and does not fix

The second phase of the India Semiconductor Mission, announced in the Union Budget for 2026-27, shifts emphasis towards equipment and materials manufacturing and towards full-stack domestic semiconductor intellectual property REPORTED. The logic is sound. A country with fabs but no local supply chain has imported its dependency rather than removed it.

The instrument, though, is imperfectly matched to the problem. India's semiconductor incentives have been built primarily around capital expenditure support: the central government committed 50 per cent fiscal support to the approved Dholera fab DURABLE, and Gujarat's state policy adds 40 per cent of the centrally approved capex assistance on top DURABLE. Capex-linked support is powerful for a fab, where capital is the dominant cost and the payback period is the central risk. It is much weaker for a materials plant, which is typically far less capital intensive per rupee of eventual revenue. Subsidising a percentage of a smaller capital base transfers less value, and it does not address what actually constrains the decision, which is demand certainty and qualification access.

Operating cost concessions are arguably more relevant here. Subsidised power at a fixed rate per unit for a decade, subsidised industrial water, and electricity duty exemption under the Gujarat Semiconductor Policy 2022-27 all bear on a chemical plant's running economics DURABLE. It should be noted honestly that the published policy language ties the most generous land subsidy specifically to a fab in Dholera Semicon City DURABLE, and whether a materials manufacturer would be treated on the same terms is not established in any public document. Similarly, the state's commitment to match approved central assistance for electronics component projects under its 2025 component policy DURABLE may or may not extend to upstream chemical supply. Those are open questions, not settled entitlements.

The demand arithmetic behind the decision

Anyone evaluating a materials plant at Dholera has to build a consumption model, and the inputs are unusually unfavourable at present. The approved fab is sized at up to 50,000 wafer starts per month on 300mm DURABLE, with commercial production guided to mid-2028 TARGET. Reporting in July 2026 indicated the line opens mostly at 90nm rather than 28nm REPORTED, and a Tata spokesperson was quoted saying the plan has always been to start with 55nm and 90nm before introducing 28nm REPORTED. The picture is not entirely settled: the Tata Sons chairman's statement in the annual report for the year ended March 2025 said the company had chosen to start its chip journey at the 28nm node DURABLE. The official node range for the project spans 28, 40, 55, 90 and 110nm DURABLE.

Mature nodes generally consume less chemistry per wafer than leading-edge ones, because chemical consumption scales roughly with the number of patterning, deposition and planarisation steps, and a 90nm flow has substantially fewer mask layers than an advanced logic flow. Back-end metallisation at mature nodes is frequently aluminium rather than copper, which changes the slurry and etch mix and removes some of the highest value consumables from the basket. In short, one mature-node fab at full ramp is a meaningful customer but may still sit below the minimum efficient scale of a dedicated resist or slurry manufacturing line.

That is why the honest reading of the timing is that the decision does not need to be taken now. Volume demand, if it arrives on the stated schedule, arrives from 2028 onwards TARGET. Holding an option is rational. It is also why, if anything is built, the industry's normal sequence would begin with blending, formulation, filtration, packaging and quality control of imported base chemistry, together with a technical service laboratory, rather than synthesis from raw feedstock. That model requires a fraction of the capital, qualifies far faster, and is how many materials suppliers have historically entered a new geography. A full synthesis plant is a much later step, and it typically waits for several customers rather than one.

What would convert an exploratory memorandum into a project

There are observable, publicly documented milestones that would distinguish a genuine project from a signing ceremony, and none of them has been reported.

• A land allotment or lease at Dholera to a named Fujifilm legal entity. The government delivery report of 31 May 2026 recorded 12 plots allotted, of which 436 acres were industrial, with industrial and commercial land recorded as readily available for further allotment DURABLE, so an allotment would be traceable.
• Environmental clearance and state pollution control board consent applications. A chemical synthesis facility in India requires environmental clearance under the EIA notification, and formulation, blending and warehousing units require consent to establish from the state pollution control board, with the applicable category depending on scale and process. Those filings are matters of public record, which makes them a reliable early indicator well before construction starts.
• Effluent and solvent handling provision. High purity chemical manufacturing generates solvent waste and acidic effluent, and the treatment infrastructure has to be designed in from the start rather than retrofitted.
• An announced capital figure and a board sanction, which is the point at which a company converts an option into a commitment.
• Any disclosed supply, qualification or joint development arrangement with a fab operator. This is the single most informative signal, because it is the one thing a state government cannot provide.

What remains unestablished

The record as it stands is thin, and it is worth stating the gaps plainly rather than inferring past them. There is no published product scope, so it is not known whether the contemplated base would cover photoresists, CMP slurries, process chemicals, specialty films, or some combination. There is no announced investment figure, no identified site within the special investment region, and no stated timeline. It has not been publicly clarified whether the facility under discussion would be a manufacturing plant at all, as against a warehousing, blending and technical service centre, which are very different propositions in capital, employment and industrial significance.

Nor is there any public confirmation of a commercial relationship with the fab that would presumably anchor demand. What exists is a memorandum with a state promotion body to explore a possibility REPORTED, signed at a moment when the anchor customer's own commercial production is guided to mid-2028 TARGET. That is a reasonable thing for a materials company to sign and a reasonable thing for a state to announce. It is not yet a plant, and treating it as one would misread both the document and the industry it describes.