India's Semiconductor Story: From Consumer to Creator

The Demand Storm India Is Stepping Into

The global semiconductor industry crossed $772 billion in 2025. By 2026, IDC projects it will hit $1.29 trillion — a number that would have seemed fantastical a decade ago. The primary driver is not smartphones or PCs anymore. It is AI.

Generative AI chips alone are projected to account for roughly half of all chip sales by 2026 — approximately $500 billion, according to Deloitte. The data centre segment is projected to reach $477 billion this year, growing to $843 billion by 2030, representing nearly half the entire semiconductor market. Hyperscaler capital expenditure — the spending by Amazon, Google, Microsoft, and Meta on compute infrastructure — crossed $100 billion in a single quarter for the first time in Q3 2025, and is forecast to reach $600 billion across 2026, a 70% year-on-year jump.

The AI chip market itself had a compound annual growth rate of 33.2% projected between 2025 and 2030. AMD CEO Lisa Su's estimate from late 2025 put the total addressable market for AI data centre chips alone at $1 trillion by 2030. Even as a conservative reference point, these numbers signal something structural: the world is not building more chips because it wants to — it is doing so because AI workloads have created demand that no existing supply architecture can comfortably satisfy.

India sits at the edge of this storm. The question is whether it gets swept up in it — or learns to navigate it.

India's Consumption Problem — and Its Scale

India currently consumes approximately $52 billion worth of semiconductors annually, per IESA estimates for FY 2024–25. That number is projected to more than double to over $103 billion by 2030, growing at roughly 13–15% per year. UBS, in an April 2025 note, put the 2030 figure at $108 billion with a 15% CAGR. IMARC Group's longer horizon projects $180 billion by 2034.

The import dependency behind this is staggering and often underappreciated. India's overall electronics import bill for FY 2024–25 was $98.6 billion — against exports of only $38.5 billion. Integrated circuit imports have grown over 2,000% between FY 2016 and FY 2024. Memory chip imports rose over 4,500% in the same period. Amplifier imports grew nearly 4,800%.

These are not the import figures of a country building its own supply chain. They are the figures of a country that has been — almost entirely — a passive consumer of chips designed, fabricated, and assembled elsewhere.

One additional dimension is often overlooked: approximately one-third of India's semiconductor imports come from China. In a world where the US-China tech decoupling is accelerating and export controls are tightening on advanced chips, this is not merely an economic vulnerability — it is a strategic one.

The Manufacturing Push: What Is Actually Being Built

India's Semicon India Programme, with an outlay of ₹76,000 crore (~$9.1 billion), has so far committed over ₹65,000 crore to approved projects, with total semiconductor investments across ten projects amounting to ₹1.52 lakh crore (~$18 billion). Gujarat alone accounts for ₹1.24 lakh crore of those commitments.

The flagship project is the Tata Electronics–PSMC fab in Dholera. Costing ₹91,000 crore (~$11 billion), this is India's first commercial semiconductor fabrication plant. At 50,000 wafers per month capacity, it is designed to serve AI, automotive, and consumer electronics segments. It is expected to create over 20,000 direct jobs. Critically, PSMC is a Taiwanese foundry — India does not have the domestic process knowledge to build this alone. This is a technology transfer arrangement, not an indigenous capability, and it is important to understand that distinction.

The Micron Technology ATMP facility in Sanand, Gujarat ($2.75 billion) was India's first significant foreign chip investment and the first to come online. Micron's facility handles assembly, testing, marking, and packaging of memory chips — not fabrication. It is a valuable piece of the value chain, but packaging is a lower-technology-intensity segment than front-end manufacturing. Micron's first India-made chips were expected by late 2024 or early 2025.

Tata Semiconductor Assembly and Test in Assam (₹27,000 crore) is another ATMP facility, targeting wire bond, flip chip, and image sensor packaging. It is expected to create over 27,000 direct and indirect jobs in a state where industrial employment at this scale is genuinely transformative.

CG Power, in partnership with Japan's Renesas and Thailand's Stars Microelectronics, is building an OSAT (Outsourced Semiconductor Assembly and Test) facility in Gujarat for ₹7,600 crore. The HCL–Foxconn JV in Uttar Pradesh was approved by the Cabinet in May 2025. A separate Tata Electronics–Intel MoU — undisclosed in financial terms — covers manufacturing and packaging of Intel products for the Indian market.

On the equipment and R&D side, Applied Materials has committed $400 million to an engineering centre, and Lam Research has invested $25 million in a semiconductor training lab targeting 60,000 Indian engineers. These investments in infrastructure and training are arguably as important as the fab announcements — they signal a longer-term bet on India's engineering ecosystem, not just its current cost advantages.

The Design Advantage India Is Not Fully Leveraging

Here is the paradox at the heart of India's semiconductor story: India produces roughly 20% of the global semiconductor design workforce. Between 125,000 and 150,000 Indian engineers work in the global semiconductor industry. Two 2nm chip design centres operate in India — one each in Noida and Bengaluru. Companies like Qualcomm, Intel, Texas Instruments, MediaTek, and AMD run significant design and R&D operations in Bengaluru, Hyderabad, Pune, and Chennai.

Yet India earns almost no revenue from this work on its own balance sheet. These engineers are employees of multinational companies — highly paid, highly skilled, making chips for products that India then imports and consumes. The IP, the product roadmaps, and the profit pools reside elsewhere.

The Design Linked Incentive (DLI) Scheme is the government's attempt to change this calculus. It offers 50% of design costs as subsidy and a 4–5% incentive on chip sales. So far, 24 chips and SoCs have been approved — covering applications in satellites, drones, IoT, AI, and LEDs. That number is modest relative to the ambition, but the quality of projects is improving.

The more interesting signal is what India's engineering colleges — particularly the IITs — are producing in terms of deep-tech chip startups.

The Startup Layer: IP-First, India-First

Early-stage semiconductor startup funding in India grew from $5 million in 2023 to $28 million in 2024, crossing $44 million by mid-2025. Over 100 startups are now active across the value chain. The aggregate market potential for this ecosystem is estimated at $150 billion by 2030, with a 24% CAGR, according to Inc42.

Mindgrove Technologies (Chennai, IIT Madras alumni) is building India's first commercial general-purpose microcontroller based on the RISC-V architecture. Their Secure IoT chip prototype targets IoT, robotics, and smart automation. What makes Mindgrove significant is that they are building product-grade silicon — chips designed to reach the market, not just research demonstrations.

InCore Semiconductor (Chennai, IIT Madras' SHAKTI programme spinout) develops RISC-V processor cores including the Azurite, targeting low-power applications like smart meters and smart cards. Their model is selling processor IP to other chip designers — a high-margin, capital-light business that does not require a fab of its own.

Saankhya Labs (Bengaluru) is perhaps India's most strategically positioned startup: they make software-defined radio chips for broadcast and defence applications, including next-generation direct-to-device broadcasting. This is a domain where geopolitical self-reliance arguments are strongest.

Signalchip (Bengaluru) has achieved what most Indian startups have not: actually fabricated wireless communication chips for 4G and 5G base stations. Taping out — completing a full chip design and sending it for fabrication — is the industry's ultimate proof-of-concept. Most startups never get there.

Netrasemi (Thiruvananthapuram) is working on AI-at-the-edge chips: processors that run inference locally without cloud connectivity. This segment will be critical for autonomous vehicles, smart cameras, healthcare devices, and industrial automation — all high-growth markets in India.

Morphing Machines (Bengaluru) raised ₹38.36 crore in a Series A in October 2025, building reconfigurable processor architectures that can dynamically adapt to different computational workloads. The company's seed round of $2.76 million came in June 2024. Reconfigurable compute is an area of genuine global R&D interest, not just an India story.

L&T Semiconductor Technologies — backed by engineering conglomerate Larsen & Toubro with $300 million committed — is targeting 15 chip designs by 2027 and has signalled a $10 billion fab ambition (though that remains contingent on several factors). Having an Indian industrial giant in this space is strategically meaningful, even if the fab timeline is speculative.

The common thread across these companies: they are building IP, not just services. That is a fundamental shift from India's IT heritage.

The Policy Architecture: What Is Enabling This

Beyond the headline incentives, several less-discussed policy moves have been meaningful.

The SEZ land reform of June 2025 reduced the minimum land requirement for semiconductor manufacturing units from 50 hectares to 10 hectares. This may sound administrative, but it was a genuine barrier for smaller OSAT and design-linked manufacturing operations. Requiring 50 hectares was a policy designed for large integrated fabs — it inadvertently excluded mid-scale investments that now have a path forward.

The India–US semiconductor partnership under the CHIPS Act's ITSI Fund, signed in September 2024, is more significant than it appears on the surface. The ITSI (International Technology Security and Innovation) Fund is the bilateral mechanism through which the US extends its CHIPS strategy globally. India's inclusion signals that Washington views India's semiconductor push as strategically compatible with its own supply chain diversification goals — not as competition.

Four new ISM approvals in Odisha, Punjab, and Andhra Pradesh — totalling ₹4,600 crore across four manufacturing units — have been announced, creating 2,034 direct skilled jobs. The geographic spread matters: semiconductor industrialisation should not only happen in Gujarat.

The PLI scheme's broader performance is worth contextualising India's chip ambitions against: PLI for Large Scale Electronics alone attracted ₹14,065 crore in investment, produced ₹9.8 lakh crore in output, and created over 1.3 lakh direct jobs. India's mobile manufacturing transformation — from 2 units in 2014 to over 300, producing ₹5.5 lakh crore in value — demonstrates that incentive-driven manufacturing can scale rapidly when policy is sustained. Semiconductors are a harder problem, but the PLI track record is not irrelevant.

What India Is Not Saying Loudly Enough

India's semiconductor ambition, as currently articulated, is heavily weighted toward ATMP (assembly, testing, marking, packaging) — the downstream, lower-value end of the chip supply chain. Of the major projects approved, only the Tata–PSMC Dholera fab is true front-end fabrication. The rest are assembly and test operations.

This is not a critique — it is a sequencing reality. You cannot skip to advanced node fabrication without first building the industrial base, the talent pipeline, the materials supply chain, and the equipment ecosystem. ATMP is the right entry point. But it should be understood as a beginning, not a destination.

The global foundry market reached a record $320 billion in 2025, and TSMC alone commands over 90% of advanced node production globally. India is not going to dislodge that concentration in a decade. What India can do — and what the current investments suggest it is doing — is become a credible alternative for mature-node fabrication (28nm and above), packaging and test, chip design, and eventually, niche advanced applications in defence and space.

IESA has set a target of 25% local value addition in electronics by FY 2025–26, rising to 40% by 2030. Getting there requires not just fabs and ATMP facilities, but a deepening of the materials, equipment, and EDA software ecosystem — areas where India currently has almost no presence.

Hyperscaler commitments to India's data centre market have exceeded $30 billion in early 2026. If India builds the right semiconductor infrastructure, a portion of that demand could be served by locally designed and assembled chips. The automotive semiconductor segment is projected to grow at 8.66% CAGR to 2031 in India. The sensors segment is projected at ~14.5% CAGR, MCUs at ~13.8%. These are not abstract figures — they describe the demand pull that could sustain a domestic chip industry over the next decade, even before India cracks the global export market.

The Honest Assessment

India's semiconductor moment is real. The investments are real, the talent is real, the policy intent appears sustained — which, given India's policy history, is the hardest variable to lock in.

But several challenges remain underappreciated in the mainstream narrative.

The Dholera fab's PSMC partnership is a technology transfer, not a domestic capability. When — and if — PSMC or TSMC decide to upgrade process nodes or consolidate operations, India will not have the indigenous knowledge to continue independently. Building that knowledge base takes a generation of engineers doing real process development work, not just operating imported equipment.

The design talent gap is inverse to what it appears. India has 20% of global semiconductor design talent — but almost all of it is employed by multinationals. Converting that human capital into Indian-owned IP requires not just DLI subsidies, but the kind of patient, long-horizon capital that India's risk ecosystem has historically been poor at providing. Semiconductor startups do not return capital in 7 years. They return it — if they return it — in 15 to 20.

And the supply chain depth is shallow. India currently has no significant presence in semiconductor equipment manufacturing, EDA software, specialty chemicals, or advanced packaging materials. These are not gaps that government incentives alone can close; they require decades of industrial policy, supplier development, and technical institution-building.

None of this negates the momentum. It contextualises it. The question is not whether India can build a semiconductor industry — it clearly can, and is beginning to. The question is whether India can build one that is deep enough, and self-sustaining enough, to matter in the geopolitics of the next semiconductor era.

India's semiconductor story is one of the most ambitious industrial pivots any democracy has attempted in the 21st century. The opening chapters are being written now — with real money, real fabs, and real engineers. What comes next depends on patience, capital, and the willingness to play a long game in an industry that does not reward short ones.