Introduction

The ongoing trade war between the United States (US) and China is driven largely by each nation’s ambition to secure a technological edge, which they view as critical to determining global leadership. Semiconductors underpin modern technology because modern electronics rely almost entirely on integrated circuits (ICs) fabricated from materials such as silicon and germanium. The advances of modern technology have thus been enabled largely by the development of ICs.

Such technology has been central to India’s economic growth, particularly in the software domain. India’s IT sector revenue grew to US$283 billion in 2024, constituting about 7.3 percent of the country’s Gross Domestic Product (GDP).^[1] India’s recent entry into semiconductor manufacturing is therefore crucial, driven by the hope of replicating its software success in hardware.

This paper offers an overview of the various facets constituting the highly complex tapestry of the global semiconductor supply chain, including a basic understanding of the technology, its manufacturing process, and the surrounding geopolitics. It concludes with an analysis of India’s semiconductor initiatives and recommendations for the upcoming Phase II of India’s Semiconductor Mission.

The Importance of Semiconductors

Since the invention of transistors at AT&T’s Bell Laboratories in 1947, the role of semiconductors in society has increased dramatically.^[2] Semiconductors are used to build transistors and diodes because, unlike conductors and insulators, their atomic structure allows them to function as either, depending on whether they are switched “on” or “off,”—a property crucial for electronics.

This was followed by the creation of ICs in 1958 by Jack Kilby at a Texas Instruments lab, which made it possible to place millions of transistors on a single chip through Very-Large Scale Integration (VLSI).^[3] The industry has since evolved in line with Moore’s Law (see Figure 1), which states that the number of transistors on an IC doubles roughly every two years, though this trend now appears to be nearing its limit.^[4] For instance, the A17 Pro chip used in Apple’s iPhone 15 contains more than 19 billion transistors.^[5]

Figure 1: Moore's Law

Source: Semianalysis^[6]

ICs are the lifeblood of modern technology, forming the foundation of nearly every electronic device today. The ability of transistors—and by extension, ICs—to store, process, sense and move signals or data makes them the basic units of electronic devices, serving as memory, sensors, communications, and power lines. As a result, semiconductors applications are pervasive, spanning established sectors such as electronics, computing, medical devices, defence, and aerospace, as well as emerging areas including artificial intelligence (AI), 5G communications, Internet of Things (IoT) devices, electric vehicles, and quantum technology. Semiconductors are therefore shaping the trajectory of modern technology. Beyond their technological indispensability, they drive global innovation. Given their dominant role in defence applications, which prompted the invention of ICs, they remain fundamental to national security, underpinning weaponry, navigation systems, communication equipment, and advanced radar systems.^[7]

In this context, it is clear why the semiconductor industry has become one of the world’s largest. In 2024, global semiconductor sales reached US$627.6 billion—the highest on record—representing a 19.1 percent increase over 2023.^[8]

Semiconductor Chips: An Overview

In this paper, the terms semiconductor, chip, semiconductor chips, or microchip are used interchangeably for ICs, unless stated otherwise. ICs are compact devices containing multiple interconnected electronic components such as transistors, diodes, resistors, and capacitors, integrated onto a single chip of semiconducting material such as Silicon.^[9] Depending on the number of transistors on a chip, ICs can be classified as Small-Scale Integration (SSI), Medium-Scale Integration (MSI), Large-Scale Integration (LSI), Very-Large Scale Integration (VLSI), and beyond.

There are several ways to classify semiconductor chips, and some of the primary methods are outlined in the following paragraphs.

By Application

Depending on the application, semiconductors are used to create a range of chips that store, process, sense, and transmit data, support displays, and manage power. These functions give rise to distinct categories of chips, each requiring unique facilities, specialisations, and manufacturing processes. The global semiconductor sales for chips, by type and as of 2022, is given in Figure 2.

Figure 2: Global Semiconductor Sales by Type (2022)

Source: SIA^[10]

Logic Chips

Logic chips are the largest category of semiconductor devices and provide the processing power in digital electronics. While they were initially limited to microprocessor units (MPUs), they are now widely used in Graphics Processing Units (GPUs), which are central to AI and machine learning.^[11] Logic chips also include microcontroller units (MCUs) and Digital Signal Processors (DSPs) chips.^[12] Their applications span smartphones, personal computers, high-performance computing (HPC), IoT devices, the automobile industry, and AI systems.

Memory Chips

Memory chips are optimised for data storage and are broadly classified as volatile or non-volatile. Volatile memory, such as Random Access Memory (RAM), holds data only while a device is turned on,^[13] with Dynamic Random Access Memory (DRAM) dominating sales in this category.

On the other hand, non-volatile memory provides long-term storage even when a device is powered off. Read-Only Memory (ROM) is designed specifically for reading data, while flash memory allows both reading and writing.^[14] NAND flash is the most widely used type and is found in Universal Serial Bus (USB) drives, memory cards, and solid-state drives (SSD).

Analog Chips

Analog chips convert continuous signals, such as temperature or pressure, into digital signals for processing by digital devices.^[15] Their applications include power management, communication devices, and military equipment such as radars and sonars. Unlike logic and memory chips, the analog market is more application-specific and typically involves smaller production volumes.^[16]

Optoelectronics, Sensors and Discrete (OSD)

Optoelectronic chips generate or manipulate light and convert it into digital signals. They include Light Emitting Diodes (LEDs), image sensors and laser diodes, which are used in display panels, cameras, and fiber optics.^[17] Sensors detect environmental changes such as heat, pressure, and acceleration and are employed across consumer electronics and industrial equipment.^[18] Discrete chips are elementary devices that function independently of larger circuits and typically perform a single electrical task, such as controlling current within an integrated circuit.

By Node Size

Table 1: Types of Transistors

Term	Elaboration
Metal Oxide Semiconductor Field-Effect Transistor (MOSFET)	Most used transistor consisting of a metal gate, insulating oxide layer, and semiconductor material.
Fin Field-Effect Transistor (FINFET)	Evolution of MOSFET consisting of a three-dimensional fin-shaped silicon structure above the substrate.
Gate All-Around Field-Effect Transistor (GAAFET)	Evolution of FinFET in which the insulting oxide and gate are wrapped around the channel material from all sides.

A technology node (or process node) is an industry label used to define and track successive generations of semiconductor technologies.^[19] Historically, a node referred to the size of key electronic components on a chip—such as the transistor gate length for logic chips or the half-pitch (half the distance between adjacent memory cells) for DRAM.^[20] In general, smaller nodes indicate more advanced chips. Recently, however, node names have become largely marketing terms, denoting generations developed with specific technologies rather than the physical dimensions of the chip.^[21]

Figure 3: MOSFET

Source: Renesas^[22]

This shift occurred because, from the 22 nm generation onwards (with exact thresholds varying by company), chip technology shifted from planar metal-oxide-semiconductor field-effect transistor (MOSFET) (see Table 1 and Figure 3) configuration to three-dimensional fin field effect transistors (FinFETs) and gate-all-around field effect transistors (GAAFETs) (see Figure 4).^[23] Consequently, gate length or half-pitch is no longer relevant in the absence of planar structures. There is currently no universal consensus among vendors on node nomenclature, which is now used primarily as a marketing term.

Figure 4: FinFET and GAAFET

Source: Renesas^[24]

As of 2023, the National Institute of Standards and Technology (NIST) has classified different node types as given in Table 2.

Table 2: Technology Nodes

Node Type	Node Length	Application
Leading-edge	Less than 5 nm for logic chips or half-pitch of less than 13 nm for memory chips	Advanced logic chips and advanced memory chips
Current-generation	Between 5nm and 28 nm	Logic, analog, radio-frequency and mixed-signal chips
Mature	Above 28nm	Logic, analog and OSD chips

Source: NIST^[25]

Therefore, while advanced nodes are crucial for leading-edge logic and memory chips, mature-node technologies remain essential for growing markets such as automobiles and 5G communications, as well as industrial and defence applications.

By Packaging

As performance demands increased, shrinking two-dimensional planar devices (i.e., MOSFET) was no longer sufficient to maintain the pace described by Moore’s Law. This drove the development of more advanced packaging and device architectures, including three-dimensional approaches such as FinFETs and GAAFETs.^[26] For instance, in memory chips, this shift is reflected in 3D NAND, where layers of memory cells are stacked on top of each other to improve storage capacity and read/write speeds, with current designs exceeding 200 layers.^[27] These devices can be produced using mature nodes (around 30nm to 50nm), making them highly cost-effective.

Another aspect of semiconductor packaging has emerged from the need for modern electronic devices, such as smartphones, to support multiple complex functions. To address this, the industry moved towards integrating components such as MPUs, GPUs, and USB controllers onto a single system-on-chip (SoC).^[28] As functionalities grew more sophisticated and costly, separate building blocks called chiplets were developed, allowing only the highest-performance blocks to use advanced nodes while others rely on mature nodes.^[29]

The high-performance demands for HPCs and AI accelerators have also led to the development of Chip on Wafer on Substrate (CoWoS) packaging which integrates logic and high-bandwidth memory chips side-by-side onto a Silicon substrate which provides superior performance and efficiency.^[30] Currently, the Taiwan Semiconductor Manufacturing Company (TSMC) is the primary manufacturer of CoWos technology.

By Semiconducting Material

The growing demand for analog power-management chips for electric vehicles and renewable energy systems has increased the use of wide-bandgap or compound semiconductors like Silicon Carbide (SiC) and Gallium Nitride (GaN).^[31] Unlike pure semiconductors like Silicon and Germanium, compound semiconductors combine elements from different groups of the periodic table. They can operate at higher temperatures and exhibit lower current leakage, making them well suited to power management and adverse environments. GaN devices, for example, are used in fast-charging consumer electronics as well as aerospace and defence systems.

As power demands rise, overheating has become a major factor contributing to the plateauing of Moore’s Law.^[32] To address this concern, novel materials such as Graphene are being explored for chip manufacturing.^[33] However, adopting these materials would require replacing existing global manufacturing infrastructure, demanding substantial investment. Even so, companies like Black Semiconductor aim to begin mass production of Graphene-based semiconductors, with a pilot facility planned in Aachen, Germany, in 2026.^[34]

The Semiconductor Manufacturing Process

Though semiconductor manufacturing and production is an incredibly complex process, involving over a thousand steps for advanced nodes, it can be broadly divided into three stages, illustrated in Figure 5.

Figure 5: Simplified Semiconductor Supply Chain

Source: CSIS^[35]

1. Design

The semiconductor manufacturing process begins at the design stage, where a blueprint of the chip’s architecture is created to optimise parameters such as cost and power consumption based on the requirements of the specific chips to be produced.^[36] As ICs became more complex, the design process evolved from hand-drawn to eventually relying completely on highly specialised software called Electronic Design Automation (EDA).^[37] Certain segments of fundamental chip design have been licensed by computer architecture companies (such as the UK’s ARM) as Intellectual Property (IP) called Core IP. These reusable components allow chip designers to avoid recreating basic elements and instead focus on innovative changes and additions.^[38] Finally, chip designs created using EDA software are then translated into manufacturing processes using Process Design Kits (PDKs), which are usually specific to technology nodes and companies.

2. Fabrication

Fabrication is simultaneously the most critical as well as complicated step in the semiconductor manufacturing process. Post design, semiconductor chips are fabricated on circular sheets of silicon (or other semiconducting materials) called wafers, which can be 8 or 12 inches in diameter. These wafers are sliced from silicon ingots consisting of 99.99 percent pure silicon, then polished and coated with thin films of conducting, isolating, or semiconducting material.^[39]

To etch out billions of tiny features (including transistors) on a chip, the wafer is first covered in a light-sensitive coating called a photoresist. It is then exposed to ultraviolet light through a photomask or “reticle” containing the circuit blueprints, a process known as photolithography.^[40] When the light hits the photoresist, it induces a chemical reaction that replicates the mask pattern on the wafer. The degraded photoresist is then removed or “etched” out, either through dry (gas-based) or wet (chemical) etching. Finally, the wafer undergoes ion implantation, during which it is bombarded with ions to tune the conducting properties of the circuit pattern.^[41]

3. Assembly, Testing, Marking, and Packaging (ATMP)

After front-end fabrication is completed, wafers are sent to back-end manufacturing facilities collectively known as Assembly, Testing, Marking and Packaging (ATMP). To extract the chips, the wafers are diced into individual units using diamond saws.^[42] The chips are then marked, tested for performance, and packaged for protection and integration with electronic devices. Packaging involves placing the chip on a substrate that directs input and output signals, along with a heat spreader to ensure cooling during operation.^[43] Different kinds of chips, such as MPUs, GPUs and memory chips, are individually packaged and mechanically assembled on a circuit board, interconnected by wires and pins. Compared with fabrication, ATMP is more labour-intensive and requires fewer complex tools and equipment.

Semiconductor Supply Chains: A History

Though the initial outsourcing of ATMP processes to the East Asian Tiger economies dates to the late 1960s, the subsequent concentration of supply chains in Asian economies is not exclusive to semiconductors. Rather, it was a natural consequence of the neoliberal economic order established by the US from the 1980s, under which large segments of manufacturing were offshored to Asian economies like China, which became the “factory of the world,” owing to cheap labour and lower manufacturing costs.^[44] In China’s case, offshoring was also part of wider efforts to integrate it into an American-led economic order in the hope of aligning Beijing’s behaviour with liberal norms.^[45] The failure of this effort over more than four decades effectively triggered the US–China trade war, with semiconductors adding further fuel to the conflict.

In semiconductors, the concentration of manufacturing in East Asian economies reflects an economic interdependence model shaped by the complex, costly and risk-intensive nature of production. This interdependence model emerged primarily as a consequence of the incentivisation for the private sector to find the most efficient and profitable means of production.^[46] As such, abundant labour and resource access, combined with favourable policies and a conducive business environment, as well as easy market access, allowed semiconductor manufacturing (particularly fabrication and ATMP) to spontaneously evolve and flourish in countries like China, Taiwan, Japan, and South Korea, which essentially led to the idea of “friendshoring” by the US.^[47]

With the rapid development in semiconductor manufacturing technologies, fabrication costs have risen substantially. In 2011, the cost for a single firm to design and prototype 28nm chips was around US$51 million, compared with US$542 million for 5 nm chips in 2020.^[48] According to TSMC, a 3 nm fabrication plant costs more than US$20 billion to build, while each 3 nm chip currently costs around US$20,000.^[49] Therefore, rising costs have further cemented semiconductor supply chains into place, fragmenting each stage and enabling domination by specific entities. This makes establishing independent supply chains extremely costly, aside from the inherent risks of such long-term investments. The established competence and proficiency of companies like TSMC has also fostered deep customer reliance built over decades, which is difficult to replicate.^[50]

While the aforementioned reasons shaped today’s semiconductor supply chains, the issue has acquired particular prominence due to a variety of factors:

1. The genesis of US-China tech rivalry can be traced back to 2017, when China’s ZTE Corporation was accused of violating its settlement with the US regarding sanctions on the sale of telecommunication equipment to Iran, a charge it pleaded guilty to in 2018.^[51] The Trump administration responded with the 2018 Export Control Reform Act, which required technologies essential for national security to be subject to export controls. This shift arguably marked the transition in US-China relations from economic engagement to strategic competition. Tensions deepened as the Chinese Government pursued ambitious indigenisation of strategic technologies under the Made in China (MIC) 2025 initiative, often through foreign acquisitions and alleged Intellectual Property (IP) theft.^[52] Subsequently, the US has employed several tools, such as the Military End User List, the Entity List and the Foreign Direct Product Rule, to target the sectors covered by the MIC 2025 policy while hampering the growth of tech giants like Huawei.^[53]
2. The COVID-19 pandemic highlighted the importance and the fragility of current semiconductor supply chains. A series of fires, earthquakes, and droughts at key manufacturing locations, coupled with rising demand for consumer electronics, exposed bottlenecks and overreliance on certain geographies within the global supply chain framework. In March 2021, a fire at a Renesas Electronics factory in Tokyo halted the worldwide flow of automotive semiconductor chips.^[54] Similarly, a drought in Taiwan in the same year forced the government to cut off the water supply to Taichung, a major chip manufacturing location.^[55] Both these instances had cascading effects across the global economy during the pandemic. Moreover, Taiwan faces recurring droughts and earthquakes that frequently disrupt semiconductor output.^[56]
3. The offshoring of manufacturing to East Asian economies has contributed to a gradual decline in US domestic manufacturing capabilities and a growing sentiment of techno-nationalism.^[57] This shift has essentially marked the globalisation era and a move towards a more multipolar world order, particularly since the inception of the second Trump administration, with a stronger emphasis on bilateral rather than multilateral engagement. Given their centrality across technology sectors, semiconductors have been among the most affected, prompting onshoring efforts and the strategic realignment of supply chains.
4. The technology war between the US and China has essentially centred in establishing dominance in AI, given its gradual infiltration into virtually every other technology. This competition is closely tied to the geopolitical issue of Taiwan. Only two companies—TSMC and Samsung—can fabricate advanced-node chips (3nm and 5 nm) required for cutting-edge GPUs, with TSMC as dominant supplier.^[58] TSMC is also the only foundry currently capable of manufacturing 1nm chips. As global competition over AI intensifies, these advanced chips, essential for GPUs and accelerators used to train sophisticated AI models, have shifted from technological assets to instruments of geopolitical leverage. This puts Taiwan in a precarious situation, caught between China’s plan for “peaceful reunification” by 2035 and US-backed restrictions on chip sales to Chinese firms.^[59]

These dynamics have prompted a broad strategic push by the US to decouple existing semiconductor supply chains from China and re-establish them within its own territory or among close allies. This has led to initiatives such as the US CHIPS Act (2022), the European Chips Act (2023), South Korea’s K-Semiconductor Belt Strategy (2021), Japan’s Semiconductor Revitalization Strategy (2021), and India’s Semiconductor Mission (2022).

The Current Global Semiconductor Supply Chain

Semiconductor Business Models

The semiconductor industry largely relied on the integrated device manufacturer (IDM) model until the 1980s, in which a single company handled most stages of chip development in-house. As design and fabrication became more expensive, the industry gradually shifted to a “fabless/foundry” model.^[60] A fabrication facility is referred to as a “fab,” while firms focused solely on design are termed “fabless.” Fabless companies contract foundries to manufacture chips.

While IDMs are still engaged in developing mature node chips, the manufacture of advanced logic chips has shifted almost completely to the fabless/foundry model, as specialisation across the supply chain allows companies to manage rising design and fabrication costs and benefit from economies of scale.

The Current Supply Chain

Though many nations involved in the semiconductor supply chain, it is dominated by five countries: the US, China, Taiwan, Japan, and South Korea—with others such as the UK, the Netherlands, Singapore and Malaysia contributing to specific segments.^[61] The semiconductor market share by company headquarters for different countries as well as the industry value added by region for the year 2024 are displayed in Figures 6 and 7.

Figure 6: Semiconductor Market Share, by Company (2024)

Source: SIA^[62]

Figure 7: Semiconductor Industry Value Added by Activity and Region (2024)

Source: SIA^[63]

Design

The design segment is dominated by the US and the EU, both in EDA and Core IP. EDA is dominated by US-based firms such as Cadence, Synopsys, and Mentor Graphics (a subsidiary of Germany-based Siemens), while Intel, Cadence, and ARM (UK) are global leaders in Core IP.^[64] The US also leads in logic design through fabless corporations like Nvidia, AMD, Apple, Qualcomm, and Broadcom. South Korea leads in memory through Samsung and SK Hynix, with US-based Micron also holding a sizeable portion of the global market.^[65]

Fabrication

Figure 8: Global Wafer Fabrication Capacity by Chip Category (2022 and 2032 Forecast)

Source: SIA^[66]

In fabrication, there are no clear overall leaders (see figure above), but different countries have established distinct niches. Taiwan leads in advanced-node logic chips, primarily due to TSMC.^[67] China produces a large proportion of mature-node chips, with foundries like the Semiconductor Manufacturing International Corporation (SMIC) and IDMs like the Yangtze Memory Technologies Corporation (YMTC) emerging as key players.^[68] The US contributes primarily to the logic chip segment through IDMs such as Intel and Texas Instruments, as well as foundries like GlobalFoundries.^[69] Currently, there are only two companies—TSMC and Samsung—that manufacture the most advanced-node chips (1nm, 3nm and 5nm), namely, TSMC and Samsung.^[70]

ATMP

Figure 9: ATMP Facilities by Region

Source: Semi^[71]

Similar to foundries in fabrication, contract ATMP manufacturers are referred to as Outsourced Semiconductor Assembly and Test (OSAT) firms. ATMP/OSAT facilities are heavily concentrated in China, Taiwan and Southeast Asian countries like Singapore, Malaysia, Vietnam, and Philippines (see figure above).^[72] However, some US companies like Micron, also maintain presence in this domain.

Materials and Equipment

Semiconductor manufacturing necessitates the employment of a wide variety of materials across the entire supply chain, including silicon wafers, photomasks, and photoresists. These include over 150 chemicals, 30 gases, and 30 minerals, encompassing acids such as sulphuric, nitric, hydrochloric, and phosphoric; gases like hydrogen, argon, neon, and helium; and minerals like aluminium, antimony, arsenic, beryllium, bismuth, gallium, and germanium.^[73] From Figure 6,Taiwan appears to be the leading supplier of materials, followed by South Korea and China. While firms from Taiwan, Japan, the US, and South Korea lead in wafer production, China is more dominant in minerals.^[74] Certain specialised materials are sourced from only a few countries, which has caused supply-chain disruptions in the past. For instance, neon supply was severely affected by the Russia-Ukraine conflict, as Ukraine accounted for about 50 percent of global output.^[75]

Chip production also requires highly sophisticated equipment for both fabrication and assembly and packaging. From Figure 6, the US appears to lead in overall equipment, followed by Japan and the EU (primarily the Netherlands). US firms like Applied Materials and Lam Research dominate in wafer-fab equipment.^[76] The Netherlands’ ASML is the world’s sole supplier of Extreme Ultraviolet (EUV) photolithography systems, essential for manufacturing advanced-node chips.^[77] Japan leads in microlithography and masking equipment, etching and cleaning tools, and assembly and testing equipment, driven by firms such as DISCO Corporation, TOWA Corporation, Accretech and Tokyo Electron.^[78]

Emerging Challenges

The growing strategic competition between the US and China has sent shockwaves across the entire semiconductor supply chain, imposing heavy financial strains on both sides and their allies. The US now faces the challenge of filling gaps in its manufacturing ecosystem that were previously addressed by Chinese friendshoring. Beyond the difficulty of attracting investment in an already strained economy, this is a long-term endeavour that will take years to bear fruit. China, meanwhile, has built a solid foundation in mature nodes but remains heavily dependent on the US and its allies for advanced-node technologies.^[79]

Export controls—semiconductors by the US and rare-earth minerals by China—are driving a gradual realignment of existing supply chains. This is particularly problematic for US allies such as Taiwan, Japan, and South Korea, all of which have major manufacturing facilities in China and are under geopolitical pressure to relocate them.  Industry leaders like Samsung and SK Hynix are gradually being forced to relocate manufacturing units from China to other nations.^[80] This landscape also poses a challenge for emerging economies like India, which are trying to enter the sector but risk being forced to pick sides, with potential geopolitical consequences. Indian firms, for instance, are already suffering from China’s export controls over rare-earth magnets, leaving them caught between US-China technological rivalry and strategic competition.^[81]

India’s Nascent Semiconductor Ecosystem

India has attempted to establish a foundation for semiconductors on multiple occasions. For instance, Bharat Electronics Limited (BEL) started manufacturing germanium-based transistors in 1962.^[82] Subsequently, in 1984, the Semiconductor Complex Limited (SCL) was established as a public-sector unit in Mohali and,^[83] over the following years, reached a level of technological sophistication close to global competitors. However, a massive fire in 1989 destroyed its fabrication facility,^[84] and the government at the time lacked the sustained strategic focus needed to develop semiconductor manufacturing capabilities. India’s first Semiconductor Policy, announced in 2007, met a similar fate.^[85]

In recent years, India has benefited from the US attempt to decouple parts of its manufacturing ecosystem from China. Tech firms such as Apple and Dell have shifted lower-end production, including mobile-phone assembly, to India.^[86] India has also become a focal point for US government efforts to establish secure supply chains in critical technologies, including semiconductors, through initiatives like the US-India Initiative on Critical and Emerging Technology (iCET) announced in May 2022.^[87] This has paved the way for India’s ambition to develop a complete semiconductor supply chain across design, fabrication and ATMP.

Renewed Push for Semiconductors

India’s renewed push for semiconductors began in December 2021 with the launch of the Production-Linked Incentive (PLI) Scheme for semiconductors, with a corpus of INR 76,000 crore (US$10 billion) under the Semicon India Programme.^[88] The scheme covers semiconductor foundries (any node), ATMP/OSAT facilities and display fabs, as well as additional sectors like micro-electrical mechanical systems (MEMS), sensors, compound semiconductors and discrete devices.^[89] Depending on the technology node and the propensity to hit PLI targets, the central government provides 50 percent of the project cost, with state governments contributing an additional 20-25 percent.^[90] This is the Indian Government’s largest commitment to semiconductor manufacturing to date and the most substantial support for any sector under the PLI scheme. Unlike PLIs for other sectors, where incentives are linked to incremental domestic revenue, funds here are provided on a “pari passu” basis, meaning they are available upfront.^[91] For approved cases, the 50 percent central subsidy is currently among the most generous globally. Additionally, the Government of India established the India Semiconductor Mission (ISM) in December 2022 as an independent nodal agency under the Digital India Corporation to guide policy execution, vet applications and attract investment.^[92]

To strengthen India’s semiconductor design ecosystem, the Design-linked incentive (DLI) scheme provided a 50 percent “product-design linked” investment by the Government of India to support enterprise launch, along with 4-6 percent “deployment-linked” incentives.^[93] The DLI supports indigenous companies engaged in semiconductor design and IP development and helps Indian start-ups acquire EDA software.

Other major schemes include the Scheme for Promotion of Manufacturing of Electronic Components and Semiconductors (SPECS), the Chip To Startup (C2S) Programme, modernisation of the SCL facility and the establishment of the India Semiconductor Research Centre (ISRC).^[94] To boost talent development, initiatives such as the SemiconIndia FutureSkills Programme and the India Semiconductor Workforce Development Programme have been created.^[95] The Government of India has also undertaken structural reforms to create a more conducive manufacturing ecosystem, including changes to the tax environment and labour policies, as well as refining its tariff, import, and customs policies.^[96]

India is actively pursuing international collaboration through multilaterals such as the Quad and through initiatives such as iCET—now renamed the Transforming the Relationship Utilizing Strategic Technology (TRUST) initiative.^[97] It has signed multiple MOUs with the US, the EU, Japan, Singapore, and the Netherlands and is exploring further opportunities with countries such as Taiwan and South Korea.^[98]

As of August 2025, ten major semiconductor projects had been approved under the ISM. These are listed in Table 3.

Table 3: Semiconductor Projects Under ISM

Indian Partner	Foreign Partner	Facility	Location	Budget
Tata Electronics	Power Semiconductor Manufacturing Corporation (PSMC) (Taiwan)	Fab (28 nm)	Dholera, Gujarat	INR 91,000 crore
Tata Semiconductor Assembly and Test Pvt Ltd (TSAT)	PSMC, Himax (Taiwan)	ATMP	Morigaon, Assam	INR 27,000 crore
Tata Projects	Micron (US)	ATMP	Sanand, Gujarat	INR 22,516 crore
CG Power	Renesas (Japan), Star Microelectronics (Thailand)	ATMP	Sanand, Gujarat	INR 7,600 crore
Kaynes Technology	LightSpeed Photonics (Singapore), AOI Electronics (Japan), Globtronics Technology (Malaysia)	ATMP	Sanand, Gujarat	INR 3,307 crore
HCL	Foxconn (Taiwan)		Jewar, Uttar Pradesh	INR 3,700 crore
SicSem Private Limited	Clas-SiC Wafer Fab Ltd. (UK)	Compound Fab (Silicon Carbide)	Bhubaneswar, Odisha	INR 2,066 crore
3D Glass Solutions Inc.		ATMP	Bhubaneswar, Odisha	INR 1,943 crore
Continental Device India Private Limited (CDIL)		Expansion of existing discrete semiconductor manufacturing facility	Mohali, Punjab	INR 117 crore
Advanced System in Package (ASIP) Technologies	APACT Co. Ltd. (South Korea)	ATMP	Andhra Pradesh	INR 468 crore

Source: The Economic Times^[99] and PIB^[100]

With Phase I of the ISM under wraps, the blueprint for Phase II is reportedly under development.

Recommendations for India’s Semiconductor Mission

1. Strategic re-alignment towards a more practical and focused approach: India’s semiconductor strategy remains scattered and diffused. As discussed earlier in this paper, semiconductor manufacturing is highly intricate, complex, expensive and particularly demanding in fabrication. No single nation can establish a complete supply chain in the short term. For instance, the Government of India is investing 50 percent of the cost of the fabrication facility in Dholera—about INR 45,000 crore (US$5.5 billion).
   
   Instead, India’s policy should primarily focus on areas that provide the greatest leverage, enabling it to become an indispensable part of the global semiconductor supply chain. India holds about 20 percent of global semiconductor design talent.
   
   The ATMP segment is also well suited to India, as it is more labour-intensive and less demanding in skill and capital. It therefore presents a more rewarding and advantageous pursuit with lower risk. India’s current investment in eight ATMP facilities is a promising step in this regard.
   
   Therefore, India needs to replace its ambitious semiconductor strategy with a more practical approach, prioritising design and ATMP over fabrication. The repeated delays at Micron’s Sanand facility illustrate the difficulty even in this segment.

2. Creating backward linkages: Since India is in the initial stages of developing a semiconductor supply chain, it is logical to invest in developing backward linkages—a strategy successfully used by Singapore in the initial stages of its semiconductor manufacturing strategy.^[107] This can be done through multiple channels, including developing spare parts for manufacturing equipment and focusing on specific chemicals and gases essential for fabrication and ATMP segments of the supply chain. For instance, neon gas, which powers lasers in the photolithography process, which accounts for 90 percent of global demand, saw major shortages after the Russia-Ukraine conflict disrupted supplies: Ukraine had previously provided about 50 percent of global output.^[108] India can offer an alternative by developing processing and refining capabilities for neon, a natural byproduct of its steel industry.
   
   The Electronics Component Manufacturing Scheme, announced in April 2025, is a good step in this direction.

3. Long-term strategy document and coordination with other technology missions: Phase I of the ISM lacked a definite long-term vision and strategy for India’s semiconductor future and did not produce a publicly available unified strategy document. A long-term strategic plan enlisting detailed milestones and timelines is not only invaluable for developing a vibrant manufacturing ecosystem, but it also creates predictability and confidence among investors, a strategy exemplified by China and other global leaders.
   
   A long-term plan would also enable the ISM to address another major shortcoming within India’s overall tech ecosystem: the lack of coordination and synergy amongst different technology missions and initiatives. Semiconductors underpin virtually every critical and emerging technology. Effective integration and interoperability between the ISM and initiatives like the National Quantum Mission and the IndiaAI Mission are therefore, essential for developing a unified tech ecosystem. Integration with other missions would also allow a rigorous assessment of the present and future domestic demands for semiconductors in other areas like AI and quantum technology.
   
   Taiwan provides a useful example. In 2003, it combined eight national laboratories into an independent non-profit institute, the National Applied Research Laboratories (NARLabs), which coordinates R&D, talent and policy across major technology domains, including semiconductors, AI, space, and biotechnology.

4. Enhancing focus on R&D: The R&D budget for the Semicon India Programme is only 2.5 percent of the total corpus.^[112] This is especially problematic given that India’s overall R&D expenditure remains low at 0.64 percent of GDP compared with global peers.^[113] R&D is integral across the semiconductor supply chain—design, fabrication and ATMP. Additionally, incorporation of AI into the design process is becoming increasingly important and will require greater investment in the future.^[114]
   
   With the gradual saturation of Moore’s Law, R&D in advanced materials like Graphene is becoming increasingly important and should also be incorporated into the ISM by integration with institutions such as the India Innovation Centre for Graphene and the India Graphene Engineering and Innovation Centre.

5. Increasing the participation of the defence sector: One of the drawbacks of the ISM has been a lacklustre participation by the defence sector. For fabrication, meeting the domestic requirements of India’s defence sector is where it can pay maximum dividends, particularly since the primary motivation in this area is achieving self-sufficiency in meeting national security requirements rather than competitive advantage. This becomes even more relevant given that defence technologies require mature-node chips, for which China is the largest manufacturer. Given China’s record of technological espionage, securing domestic production for defence needs is a logical strategy. It would also help establish a foundational fabrication capability in India that could be expanded over time. Therefore, a greater involvement of defence establishments and PSUs such as BEL in Phase II of the ISM is certainly warranted. Furthermore, creating a dedicated foundry for defence chips would be a practical first step towards this goal.
6. Re-assessing Location: Location is critical when building fabrication or ATMP facilities. Semiconductor manufacturing depends heavily on infrastructure, connectivity, electricity and water availability. For instance, a TSMC fab requires about 34 million litres of pure water daily, with each chip demanding almost 30 litres.^[116] Semiconductor manufacturing equipment is extremely fragile, and even a single bump during transportation can destroy it. While it is sensible to build facilities in areas like Gujarat and the South Indian states due to the abundance of infrastructure, transport, and port facilities, it does not make sense when it comes to states like Assam, which are currently lacking in this domain, not to mention the fact that maritime transportation of equipment is not possible here whatsoever.
   
   Developing more Special Economic Zones (SEZs) and industrial parks would also be beneficial, particularly in North India where these are lacking. These can help India develop clusters tailored specifically to meet the specialised infrastructure and supply requirements of the semiconductor sector.

7. Forging New Collaborations: With the gradual emergence of a multipolar world order, India must explore opportunities with new partners. Currently, India has limited partnerships with prominent and emerging players in Southeast Asia, such as Singapore, Malaysia, Indonesia, Thailand and Vietnam. Given their increasing importance in global semiconductor supply chains, India should pursue further collaboration with these nations through MOUs, forums for sharing best practices, creating R&D and investment funds, and new multilateral mechanisms.

Conclusion: From Ambition to Reality

With US-China strategic competition intensifying, India finds itself at a crossroads in its semiconductor manufacturing strategy. Given the obvious criticality of the technology and its impact across multiple sectors, it may appear sensible to be ambitious and go all in. However, a more patient and balanced approach would be wiser.

India must leverage the geopolitical rift between the US and China while aligning its policy of strategic autonomy, allowing it to draw on the strengths of both sides and gradually position itself as an indispensable part of both supply chains. Its strategy must prioritise what benefits its growing economy and labour force the most rather than succumb to geopolitical pressure from either side.

---

Prateek Tripathi is Associate Fellow, Centre for Security, Strategy, and Technology, ORF.

---

All views expressed in this publication are solely those of the author, and do not represent the Observer Research Foundation, either in its entirety or its officials and personnel.

Endnotes

• Developing and Emerging Economies
• China
• India
• USA and Canada
• Advanced node and mature node chips
• ATMP / OSAT facilities
• Chip fabrication and foundries
• Global semiconductor industry
• India Semiconductor Mission
• Semiconductor design ecosystem
• Semiconductor geopolitics
• Semiconductor manufacturing
• Semiconductor supply chain
• US–China technology rivalry

The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.