India’s semiconductor story is moving from policy ambition to operating reality.

SEMICON India 2026 closed on 19 September with more than 600 exhibitors, around 300 international companies, representation from 52 countries and 56 announced MoUs or strategic initiatives across design, manufacturing, packaging, materials, equipment, R&D and talent. Applied Materials announced its India Vision 2035, involving USD 5 billion of investment over the next decade and a tenfold expansion of India-based supply-chain capacity, while Lam Research announced approximately ₹10,000 crore for its first silicon-component manufacturing and vertical ingot-processing facility in India. (Press Information Bureau)

The commercial opportunity is obvious. The international tax question is less visible.

When a global semiconductor group places engineers, fabs, validation facilities, process-development teams, manufacturing assets and suppliers in India, deciding where the intellectual property legally sits is only the beginning.

The more important question is:

Which entity is actually creating the value, controlling the risks and therefore entitled to the profit?

That decision should be made before the operating model becomes too difficult to change.

Semiconductor Investment Is Different From Ordinary Manufacturing

A semiconductor operation rarely consists of one simple manufacturing function.

A global group may have chip architecture in the US, process technology in Japan, equipment engineering in India, wafer fabrication in another country, packaging and testing in India or Southeast Asia, and customers across multiple markets.

The value chain can involve patents, process know-how, proprietary software, equipment design, manufacturing recipes, yield optimisation, supplier engineering and highly specialised technical teams.

India is deliberately trying to capture more of this value chain. The India Semiconductor Mission currently offers fiscal support of up to 50% of project cost for qualifying semiconductor fabs and 50% of capital expenditure for specified compound semiconductor, silicon photonics, sensor and ATMP/OSAT projects. The Design Linked Incentive programme separately supports semiconductor design and IP-related activity. (India Semiconductor Mission)

Those incentives can influence where a group invests.

But they do not answer the transfer pricing question of how much profit the Indian entity should earn.

Legal Ownership of IP Does Not Automatically Decide the Return

Assume a US semiconductor company establishes an Indian subsidiary.

The US parent legally owns the core patents and initially determines the global product roadmap. India begins as an engineering support centre and receives a cost-plus return.

Five years later, the Indian team is no longer merely executing instructions.

It develops improvements to manufacturing equipment, solves process problems, determines aspects of product design, supervises supplier engineering and makes decisions that materially influence yield, reliability or commercialisation.

Can the group continue saying that all economically significant IP belongs entirely to the US merely because the patents remain registered there?

Under the OECD Transfer Pricing Guidelines, legal ownership is an important starting point, but legal ownership alone does not determine entitlement to the returns from an intangible. Transfer pricing must examine which entities perform or control important functions relating to the development, enhancement, maintenance, protection and exploitation of intangibles, commonly described as DEMPE functions. (OECD)

The UN Transfer Pricing Manual follows the same underlying arm’s-length logic. It specifically emphasises identifying who contributes to intangible value and how those contributors should be remunerated. (Digital Library)

This principle is especially important in semiconductor businesses because engineering know-how can evolve continuously and may never appear neatly as a registered patent.

A Cost-Plus Model Can Be Correct Initially and Wrong Later

There is nothing inherently inappropriate about remunerating an Indian semiconductor R&D or engineering company on a cost-plus basis.

If the overseas principal determines the research programme, funds development, controls the major technical and commercial risks, owns the relevant technology and genuinely supervises the Indian team's activities, a routine or limited-risk characterisation may be supportable.

The problem arises when actual conduct changes.

Applied Materials, for example, already describes its Indian capabilities as including product development, R&D, simulation, testing, validation, process engineering and component qualification. Its Bengaluru collaborative-engineering model is designed to bring engineers, suppliers and academic institutions together to accelerate development and commercialisation of semiconductor equipment technologies. (Applied Materials)

That does not tell us anything about Applied Materials' actual transfer pricing model, which is not publicly available.

But it illustrates the broader issue perfectly.

As India moves deeper into sophisticated engineering and innovation, multinational groups should periodically ask whether an entity originally described as a routine development centre is still performing only routine functions.

The operating model can evolve much faster than the transfer pricing policy.

Manufacturing Raises a Different Set of Questions

Now consider an Indian fabrication or packaging entity.

A genuine contract manufacturer may receive technology, specifications and materials from an overseas principal, perform defined manufacturing activities and earn a benchmarked return while the principal bears market, inventory, capacity and technology risks.

But the analysis changes if India owns major production assets, selects suppliers, develops manufacturing know-how, controls capacity planning, bears inventory and yield risks, or contributes proprietary process improvements.

In those circumstances, simply applying a routine contract-manufacturing margin may no longer capture the economic contribution.

India’s Income-tax Act, 2025 requires income from international transactions to be determined having regard to the arm’s-length price, and section 165 requires the most appropriate transfer pricing method to be selected having regard to the nature of the transaction, the associated enterprises and functions performed. (Etds)

This is where functional analysis matters more than labels such as “contract manufacturer”, “R&D centre” or “support entity”.

When Might Profit Split Become Relevant?

One-sided methods such as TNMM are highly useful when one party performs relatively routine functions and reliable comparables exist.

But semiconductor businesses can become difficult when both sides of a transaction make unique and valuable contributions.

Suppose a US parent owns foundational patents and controls global commercial strategy, while its Indian subsidiary develops important process technology, employs specialised engineers and exercises meaningful control over significant technical risks.

In such a case, benchmarking only the Indian entity against ordinary engineering-service companies may become increasingly difficult.

The OECD identifies circumstances such as unique and valuable contributions by both parties, highly integrated operations, or shared economically significant risks as factors that may point toward a transactional profit split method. That does not mean every sophisticated semiconductor operation requires profit split; the method must still follow the actual facts. (OECD)

The important point is that the transfer pricing method should evolve with the value chain.

Incentives and Transfer Pricing Should Not Be Confused

Another issue deserves careful attention.

Government support can make India economically attractive for manufacturing or R&D, but a subsidy does not automatically determine the arm’s-length remuneration of the Indian company.

The group still needs to decide who owns the assets, who controls the development programme, which entity bears cost overruns and technology failure, how government incentives are reflected economically and whether those incentives affect the appropriate cost base or profitability analysis.

Those questions are fact-specific.

A semiconductor project should therefore not have one team designing the incentive structure and another designing transfer pricing months later.

The two analyses should be commercially aligned, even though they address different legal questions.

The Structure Also Determines Future Cash Flows

IP ownership and transfer pricing eventually affect how cash moves across borders.

An Indian manufacturing company may pay royalties for overseas technology.

A foreign parent may reimburse Indian R&D.

An Indian entity may license locally developed technology to overseas affiliates.

Contract manufacturing, engineering services, equipment leasing, financing and guarantee arrangements can create additional related-party flows.

Each payment can have its own transfer pricing, withholding tax and treaty implications.

For multinational groups, this means the correct semiconductor structure should be designed around the entire lifecycle of the investment, not simply the first capital injection.

Where India-US, India-Japan, India-Singapore or other treaty relationships are involved, royalty, service and financing flows should be modelled under the relevant bilateral treaty and domestic law before intercompany agreements are finalised.

What I Would Decide Before the First Major Investment

The most important exercise is not choosing the markup.

It is deciding the operating architecture.

Which entity owns the pre-existing technology? Who funds new development? Who decides which R&D projects continue or stop? Who controls budgets? Where do the senior engineers sit? Who bears manufacturing failure and capacity risk? Who owns improvements developed in India? Can those improvements be used globally? And how will India be compensated if its role becomes more valuable over time?

Once those answers are clear, the intercompany agreements, IP arrangements, transfer pricing method and benchmarking become much easier to defend.

If they are not clear, no amount of benchmarking can repair the underlying ambiguity.

Closing Perspective

India’s semiconductor opportunity is not simply about attracting factories.

The larger opportunity is to participate in design, engineering, R&D, equipment development, packaging, manufacturing and global supply chains.

That evolution is positive for India and for multinational groups seeking greater geographic resilience.

But it also means that multinational tax structures designed for a simple captive-service model may gradually become outdated.

A group can legally own a patent in one country, manufacture in another and develop improvements somewhere else.

Transfer pricing asks a different question:

Where is the economic value actually being created, controlled and put at risk?

For global semiconductor groups investing in India, answering that question early may be just as important as choosing the fab location or negotiating the incentive package.

Because once high-value people, technology and decision-making begin moving into a jurisdiction, the profit model eventually has to follow the business model.

Research basis

This article has been prepared using official SEMICON India 2026 and India Semiconductor Mission materials, Applied Materials public disclosures, the Income-tax Act, 2025 as amended through Finance Act 2026, the OECD Transfer Pricing Guidelines 2022 and related guidance on intangibles, and the UN Practical Manual on Transfer Pricing. Research cut-off: 21 September 2026. (Press Information Bureau)