India Has ₹1 Lakh Crore in Chip Pledges. Will the Factories Actually Follow?

Semicon 2.0 has drawn about ₹1 lakh crore in investment commitments and Odisha plans a chip hub at Naraj. Semicon 1.0's record shows why pledges need tracking.

On **17 September 2026**, at the SEMICON India 2026 conference in New Delhi, Electronics and IT Minister **Ashwini Vaishnaw** said the second phase of India's chip programme had already drawn investment commitments of about **$11–12 billion, or roughly ₹1 lakh crore**. The money, he said, would be spread over two to three years.

The commitments cover much more than chip factories. They span semiconductor equipment, materials, speciality gases and chemicals, assembly-and-test units, substrates and even wafers. The government expects Semicon 2.0 to create at least one lakh jobs, a figure the minister called conservative.

A day later, at the same event, Odisha Chief Minister **Mohan Charan Majhi** announced an **870-acre "Odisha Silicon Valley"** at Naraj near Cuttack, meant to bring semiconductor, electronics and IT projects together in one hub. The state said it had received semiconductor investment proposals worth about **₹23,600 crore**, with potential for 6,100 jobs.

On paper, this is a sharp step up in India's ambition in an industry dominated by Taiwan, South Korea, the United States, Japan and China. The harder question is how much of it becomes working factories, and when.

The Union Cabinet approved Semicon 2.0 on **15 July 2026** with an outlay of **₹1,27,500 crore**, and the Ministry of Electronics and IT notified its rules on 31 August. The first phase, approved in December 2021 with ₹76,000 crore, focused on fabs (chip fabrication plants) and packaging units. The second phase keeps those but adds chip design, machines and materials, research, and talent, for six pillars in all.

Two numbers are easily confused. The **₹1.27 lakh crore** is what the government plans to spend on incentives. The **₹1 lakh crore** is what private companies say they intend to invest. One is a budget line; the other is a pledge.

The scheme will accept applications for three years, and projects generally get up to six years to complete. A mid-term review is due after three years. The India Semiconductor Mission (ISM), the unit under the IT ministry that runs the programme, will appraise applications, recommend approvals and oversee delivery.

The incentives are also less generous than before. Semicon 1.0 paid a uniform 50% of project cost for fabs and packaging plants. Under Semicon 2.0, silicon fabs get 40%, other fabs 35%, advanced packaging 35% and conventional packaging 25%. The government is betting that the early risk has been absorbed and that investors now need less of a push.

An investment commitment is a statement of intent. Some are backed by signed agreements and land allotments; others are closer to a letter of interest made at a crowded conference. None of them is the same as a plant that ships chips.

The minister's figure did not come with a list of companies or a split between binding agreements and softer pledges. Nor is it clear how much of it overlaps with Odisha's ₹23,600 crore in proposals, or with announcements by other states at the same event. Until that detail is public, ₹1 lakh crore is best read as a measure of interest, not of capital on the ground.

Chip plants are among the hardest industrial projects anywhere. A modern fab needs ultra-clean rooms, imported tools costing thousands of crores, a steady supply of speciality gases and chemicals, uninterrupted power and huge volumes of purified water. Even global leaders stumble: Intel put its planned fab in Germany on hold and later dropped it, and TSMC's first Arizona plant started later than first planned. That is not a reason to dismiss India's push. It is a reason to track it over years rather than celebrate it on announcement day.

The best guide to Semicon 2.0 is the record of its predecessor.

Under the first phase, the government approved **12 manufacturing units** with cumulative investment of over **₹1.64 lakh crore** across Gujarat, Assam, Uttar Pradesh, Odisha, Punjab and Andhra Pradesh. They include one silicon fab, one silicon carbide fab, one gallium nitride micro-LED display fab and nine assembly, testing, marking and packaging (ATMP/OSAT) units, the plants that take finished wafers, cut them into chips, encase them and test them.

By September 2026, the government said **five** of the 12 had started commercial production. **Micron's** ₹22,516-crore memory assembly and test plant at Sanand, Gujarat, formally opened on 28 February 2026 and shipped its first made-in-India memory modules. Kaynes Semicon and CG Semi, also at Sanand, are producing, and the government counts two smaller units among the five.

All five are packaging and assembly plants. None of the three fabs is making chips yet.

The largest project, **Tata Electronics' ₹91,526-crore fab at Dholera** in Gujarat, built with Taiwan's Powerchip Semiconductor Manufacturing Corporation (PSMC), shows how timelines move. At the groundbreaking in March 2024, the minister said the first chip would come out by December 2026. In July 2026, the target for commercial production was put at mid-2028, with trial output planned for late 2026. Reports in July said the plant would open on older 90nm and 55nm processes before moving to 28nm; Tata says its initial mix includes 28nm and that the project is on track against its commitments to the government.

Tata's ₹27,120-crore packaging plant at Jagiroad in Assam's Morigaon district, designed to handle 48 million chips a day, is still under construction.

The first-phase record also includes ventures that fell away before a brick was laid.

In February 2022, **Foxconn and Vedanta** announced a $19.5 billion joint venture for chips and displays in Gujarat. Talks with a technology partner stalled, the government raised questions about the cost estimates submitted for subsidy, and Foxconn pulled out in **July 2023**. The two other fab proposals from that first round also stalled. The ISMC consortium, which counted Israel's Tower Semiconductor as its technology partner, froze its plan while Tower's proposed takeover by Intel was pending. The proposal from Singapore-based IGSS Ventures did not clear the government's advisory committee.

Later, Maharashtra approved a **$10 billion Adani–Tower** fab in September 2024. It never received central approval, and by May 2025 both sides had confirmed the partnership was over.

So for the fab proposals of 2022, the conversion rate from announcement to working plant is zero. The fabs now being built came from a later wave of approvals, starting with Tata in February 2024. That is progress, but it is also the context every new headline number needs.

Odisha is not starting from scratch. In August 2025, the Union Cabinet approved two ISM projects at Info Valley in Bhubaneswar worth about ₹4,009 crore. **SiCSem**, working with Britain's Clas-SiC Wafer Fab, is building what the government calls India's first commercial compound semiconductor fab, making silicon carbide devices used in electric vehicles, fast chargers, railways, solar inverters and defence. **3D Glass Solutions** is setting up an advanced packaging and glass substrate unit. At SEMICON India 2026, the state also signed a letter of intent with QuadQuantum for a silicon carbide substrate wafer plant.

The state's incentives are among the richest in the country. Under a policy amendment approved in August 2026, Odisha will add **25% of eligible capital cost** on top of central support for ISM-approved projects, now extended to equipment, speciality gases, chemicals and materials. The money is to be released alongside the Centre's payments and linked to project milestones. For an ISM-backed project, combined public support can therefore reach well over half the cost of a plant.

The Naraj site sits near the Mahanadi river, on the Cuttack–Bhubaneswar corridor, with road links to the Paradip and Dhamra ports. The chief minister cited the river as a key advantage for industrial water supply.

Three practical tests will decide whether the hub fills up.

**Water.** A large fab can use around 10 million gallons, or about 38 million litres, of ultrapure water a day, and producing ultrapure water takes roughly 1.4 to 1.6 litres of ordinary water for every litre made. The Mahanadi also supplies drinking water to Cuttack and irrigation downstream. A public water-balance study, especially for the dry months, would settle doubts faster than assurances.

**Power.** Fabs draw tens of megawatts around the clock, and even a brief voltage dip can ruin wafers mid-process. Dedicated feeders and backup supply will matter more than the state's headline generation capacity.

**People.** The minister said in August that India had met its target of training 85,000 semiconductor engineers in four years instead of ten, and set a new goal of one lakh more. Odisha has strong institutions, including IIT Bhubaneswar and NIT Rourkela, and its policy funds skilling, internships and engineer relocation. But running a fab takes years of hands-on experience. Tata, for instance, is training around 300 Indian graduates at PSMC's plants in Taiwan.

Chip incentives in India come in layers. The core is ISM's capital support, a share of project cost paid as the plant is built. States add top-ups: Odisha's 25%, and separate semiconductor policies in states such as Gujarat and Uttar Pradesh. Land, power tariffs, stamp duty and water charges are often discounted further.

Around this sits wider electronics policy. The production-linked incentive (PLI) scheme for large-scale electronics, launched in 2020, rewards companies for higher output, mainly in phone assembly. The Electronics Component Manufacturing Scheme, approved in 2025, targets parts such as circuit boards and camera and display modules. The Design Linked Incentive scheme supports chip design firms; 24 chip design projects were approved under the first phase.

The stakes are large. Industry estimates put India's chip market at $45–50 billion in 2024–25, rising to $100–110 billion by 2030. A NITI Aayog roadmap says India imports 90–95% of the chips it uses, spent nearly $150 billion on semiconductor imports between 2016–17 and 2024–25, and could face an annual import bill of around $240 billion by 2035 if current trends hold.

Supporters of the programme have a strong case. In under five years, India has gone from no commercial chip packaging to five operating units, including a Micron plant that will have one of the world's largest single-floor assembly and test cleanrooms once fully ramped up. Approvals that once dragged on for years now take months; the minister has cited cases cleared in 90 to 200 days. Global companies are also looking for locations outside East Asia, which gives India an opening it did not have a decade ago.

Sceptics have a strong case too. Packaging is the lower-value end of chipmaking. The fabs, where most of the value and know-how sit, are still years away. The flagship fab has slipped by about 18 months, and every fab proposal from the first round collapsed. With the Centre paying up to half of project cost under the first phase and states adding more, taxpayers carry a large share of the risk.

Both readings can be true at once. The test is not which side wins the argument today, but what the numbers show in 2028 and 2030.

**What is the difference between a fab and a packaging unit?**

A fab makes the chip itself, building circuits onto silicon wafers through hundreds of precise steps. A packaging unit takes finished wafers, cuts them into individual chips, seals and tests them. Fabs cost many times more and need far deeper expertise.

**Is the ₹1 lakh crore government money?**

No. It is what private companies say they plan to invest. The government's own planned spending under Semicon 2.0 is ₹1,27,500 crore, paid out as incentives over several years.

**Will India make the most advanced chips?**

Not soon. Dholera is designed for mature processes, 28nm and older, used in cars, appliances, power electronics and telecom gear. Cutting-edge chips for phones and AI remain concentrated in Taiwan and South Korea.

**Why build in Odisha?**

The state already hosts two ISM projects, offers one of the richest state top-ups, and has ports, minerals and engineering colleges. Whether it can supply water, power and skilled staff at fab scale is still to be shown.

**From pledge to approval.** How many Semicon 2.0 commitments turn into ISM applications, and then into Cabinet approvals, within the three-year application window. Approvals are public; the ratio of approvals to announced commitments is the first real test.

**The first fab.** Whether Dholera begins trial production by the end of 2026 and commercial output by mid-2028, and at which technology node.

**Odisha's hub.** When land at Naraj is formally allotted, when the first unit breaks ground, and whether the state publishes a water and power plan. SiCSem's progress at Info Valley, where groundbreaking took place in November 2025, is the nearest benchmark.

**Money out the door.** How much central and state support has actually been disbursed against the outlays announced. The Semicon 2.0 mid-term review, due around 2029, should report this.

**Jobs.** Whether the promised one lakh jobs show up in employment data, not just in speeches.

₹1 lakh crore in pledges is real news and a sign that global suppliers are looking at India seriously. But it is an input, not an outcome. Semicon 1.0's record, with five packaging plants working, no fab yet producing, a flagship delayed and a string of early ventures abandoned, shows both how far India has come and why announcements need follow-through. The number that matters will be the share of these pledges that turns into operating plants by the end of the decade.

**Note on perspective**: {{Based on official statements, Cabinet and ministry notifications, state government releases and company disclosures as of 25 September 2026. The analysis focuses on whether investment pledges convert into operating capacity, weighing real progress against missed timelines.}}