By May 2026, India had approved twelve semiconductor projects under the ISM with cumulative investment of ₹1.64 lakh crore. Gujarat holds six. Telangana, India's densest semiconductor design cluster, holds none, despite shortlist presence on the first wave's four projects. UP offered the highest financial framework of any state and secured its first unit only in May 2025. The states that won did not offer the most. They resolved the most before the investor arrived. What separates a state that converts shortlist presence into a commissioned project from one that competes on paper?
The semiconductor allocation pattern is the sharpest available illustration of how India's federal industrial policy operates in practice; and why the states that win are not necessarily those that offer the most, but those that resolve the most before the investor has to ask.
Gujarat's institutional sequence is specific and instructive. Gujarat launched India's first state-level semiconductor policy in 2022, before ISM's first approval cycle had concluded and before any other state had a comparable framework in place. The policy was not a statement of intent; it was an operational package. Land subsidy at 75% on the first 200 acres in Dholera Semicon City. Electricity discount of ₹2 per unit for ten years, with full exemption from electricity duty. Water supply at ₹12 per cubic metre for the first five years. 100% stamp duty and registration fee refund. A single-window clearance mechanism that processed all approvals under one institutional authority. And additional capital assistance at 40% of the capex support provided by the central government under ISM.
The institutional sequence has compounded. By May 2026, Gujarat holds six of the twelve ISM-approved semiconductor projects. The 5 May 2026 Cabinet additions brought Crystal Matrix Limited's compound semiconductor fabrication facility at Dholera (₹3,068 crore, integrated mini/micro-LED manufacturing with GaN foundry services on six-inch wafers) and Suchi Semicon's OSAT plant at Surat (₹868 crore, scaling to over a billion chips per year for power electronics and analogue ICs). These were the last two units approved under the first phase of ISM, which closed with twelve projects across multiple states and ₹1.64 lakh crore in cumulative investment.
The critical institutional insight is not the generosity of the package. It is that the package was assembled and resolved before the first investor arrived. When Micron evaluated Sanand, it found 93 acres of GIDC land allotted within days of the MoU, utility infrastructure being readied in parallel with construction planning, and a state government that had already worked through the environmental, power, and water approvals that would have taken other states months to process sequentially. The MoU was signed in June 2023 during the Prime Minister's state visit to the United States. Groundbreaking happened in September 2023. Commercial production commenced on 28 February 2026; the first memory modules shipped to Dell Technologies. Kaynes Semicon at Sanand followed: foundation to commercial production in fourteen months, inaugurated on 31 March 2026 as the country's second commercial semiconductor production unit. By the Electronics and IT Minister's stated roadmap, delivered at the Kaynes inauguration, four plants will reach commercial production in 2026, two more in 2027, and Tata Electronics' Dholera fab, India's first commercial-scale fabrication facility, is targeted to be operational by 2028.
Gujarat also brought an industrial advantage that was not manufactured for the semiconductor programme but had been built over decades. The Dahej chemical cluster is a material supplier to semiconductor manufacturing. Gujarat hosts India's only operational Petroleum, Chemicals, and Petrochemicals Investment Region. The state commissioned India's first desalination plant at Dahej to provide semiconductor-grade water. The Dholera Special Investment Region, conceived as an industrial corridor node, provided the land scale and infrastructure planning that a 200-acre fab requires. These were not incentive announcements. They were industrial infrastructure facts on the ground.
Telangana's competitive position operates on a different logic, and its strength is one that no other state can replicate. Hyderabad is India's pre-eminent semiconductor design centre, with virtually every major global semiconductor firm operating design centres there. When an ATMP facility needs to iterate on packaging design with the chip designer, the iteration happens in a twenty-minute drive when both are co-located. Hyderabad has the design talent for this. Sanand does not. It is an engineering productivity advantage that compounds over every production cycle.
Yet Telangana was on the shortlist for four major projects; Micron's ATMP, Kaynes Semicon's OSAT, HCL-Foxconn's OSAT, and ASIP's advanced packaging facility; and all four located elsewhere. By May 2026, with the first phase of ISM closed, Telangana holds zero approved ISM projects. The institutional reading reflects the interplay of centralised incentive architecture, political signalling, and the timing of state-level engagement relative to ISM decision-making cycles. Gujarat moved first with a resolved package. Telangana had the design cluster but had not yet assembled the machinery to convert shortlist presence into allocation outcome.
The contrast with states that competed on paper but lost on the ground is equally instructive. Uttar Pradesh offered a more generous financial framework than Gujarat; a 50% additional capital subsidy on top of the central ISM subsidy against Gujarat's 40%, plus 100% electricity duty exemption and a dual-grid power architecture. Yet UP attracted no major semiconductor unit in the first five ISM approvals; the HCL-Foxconn JV near Jewar was approved by Cabinet only in May 2025, as the sixth ISM unit, by which time Gujarat had already secured four. The institutional lesson is precise: semiconductor investors do not evaluate incentive ceilings in isolation. They evaluate ground-level readiness; the availability of engineering talent at scale, the quality of utility infrastructure, the processing speed of state approvals, the presence of adjacent industries that supply specialty chemicals, gases, and precision components, and the demonstrated institutional capacity to deliver what has been promised within the committed timeline. UP offered the higher numbers on paper. Gujarat offered the most resolved industrial base on the ground. The investment followed the ecosystem, not the number.
The fiscal paradox deepens the institutional lesson. Comptroller and Auditor General data for 2023-24 shows UP as one of the stronger fiscal performers among Indian states, with real revenue growth exceeding 6%, supported by diversified tax and excise sources. A state that is outperforming Karnataka fiscally but did not feature in the first five ISM approvals reveals that the determinants of investment geography operate on a different institutional logic than the determinants of revenue performance; UP eventually secured HCL-Foxconn in May 2025, but as the sixth unit, after the first-wave architecture had already settled.
Odisha's trajectory carries a different institutional signature. The state moved late but moved decisively. On 19 April 2026, Odisha became the first state in India to host both a compound semiconductor fabrication unit and an advanced 3D glass-substrate packaging facility, with the groundbreaking of 3D Glass Solutions' ₹1,943 crore plant at Info Valley, Bhubaneswar. The facility is the country's first glass-substrate-based advanced semiconductor packaging unit, targeted at AI, 5G, defence, and data-centre applications, with commercial production scheduled for 2028. The institutional architecture behind Odisha's emergence is not built on the longest incentive history; it is built on the most recent state-level decisions about land allocation, single-window clearance velocity, and explicit Centre-State alignment around the project category each side wanted to bring. Odisha's case is the institutional inverse of Telangana's: less design density, more institutional readiness to convert a central decision into ground commitment.
Karnataka signed an MoU with a consortium that included Tower Semiconductor for a proposed $3 billion investment. The MoU did not materialise into an approved ISM project. The institutional dynamics behind this non-conversion are specific: an MoU at a summit is not an ISM approval. Converting the MoU into a project requires alignment with ISM's evaluation criteria, state-level co-investment structuring that matches the central incentive architecture, and site-specific resolution of land, utilities, and environmental clearances. The states that treat the MoU as the outcome, rather than as the starting point of a regulatory journey that requires sustained engagement at every level, consistently discover that the summit creates the moment but the machinery creates the project.
Rajasthan's entry on 15 May 2026 illustrates a different institutional pathway entirely. The Sahasra Semiconductors ATMP/OSAT unit at Bhiwadi, virtually inaugurated by the Union Minister for Electronics and IT, is the first semiconductor unit in the state, but it sits outside the ISM. The instrument is the Scheme for Promotion of Manufacturing of Electronic Components and Semiconductors (SPECS); the investment is ₹150 crore; the annual packaging capacity begins at 60 million units and is targeted to scale to 400-600 million over two to three years. Rajasthan released its Semiconductor Policy only in March 2026, after the first phase of ISM had effectively closed. The state has, therefore, entered the architecture through the SPECS route rather than the ISM route; the instrument is different, the scale is smaller, but the institutional signal is precise. States that arrive after the framework has been written enter through whichever instrument the architecture still has open; the choice of scheme is not strategic, it is residual.
Three structural observations emerge from the semiconductor allocation pattern.
First, the states that win industrial allocation in India are those that present resolved packages, not menus. A menu says: "here is our land policy, here is our power tariff, here is our water rate, here is our stamp duty exemption; apply separately for each." A resolved package says: "here is 93 acres, allotted; here is the power connection, sanctioned; here is the water supply, contracted; here is the environmental clearance, obtained; here is the state co-investment, committed. When can you break ground?" The difference between the two is not policy design. It is institutional execution velocity.
Second, industrial density matters more than incentive quantum. Gujarat's six semiconductor projects are not the result of the highest subsidies in India. They are the result of the densest industrial base; chemicals, logistics, established manufacturing infrastructure, and institutional credibility built over decades of industrial development. Telangana's design cluster is a different kind of density, but equally consequential for the segments where design proximity determines manufacturing productivity. The states that will capture the second wave of ISM investment are those that build genuine industrial depth, not those that announce the highest percentage.
Third, the geography of India's semiconductor future is still being determined. ISM 1.0 concentrated first-phase fabrication in Gujarat, with ATMP and packaging distributed across Gujarat, Assam, UP, and Odisha; twelve projects, ₹1.64 lakh crore in cumulative investment, four commercial production units expected to be operational by the end of 2026, two more in 2027, the Dholera fab in 2028. The first phase has closed. ISM 2.0, focused on equipment and materials manufacturing, advanced design capability, and indigenous IP, will determine whether the second wave follows the political logic of the first or the technical logic of where design talent and industrial density are already concentrated.
India's semiconductor programme is not static. Micron's Sanand facility converts imported DRAM and NAND wafers into finished memory products. But it is not planned to host High Bandwidth Memory manufacturing, the segment where global value is now concentrating because of AI data centre demand. India built the ATMP facility, but the highest-value segment in memory packaging is not yet part of the Indian production mandate. This reinforces the structural pattern visible across India's incentive architecture: domestic manufacturing depth remains one layer behind where global value is migrating.
ISM 2.0 has explicitly signalled that advanced memory packaging, including HBM, and the upstream equipment-and-materials layer will be given precedence in the next allocation cycle. This means the second wave of investments will require even more sophisticated design-manufacturing proximity than the first. The second-wave geography, if it follows the technical logic rather than the first-wave political logic, favours the states that built design density over those that built incentive packages.
India's semiconductor allocation is not a story about which state offered the most money. It is a story about which states built administrative systems that semiconductor investors could evaluate, validate, and commit to within the timelines that global capital deployment demands. Gujarat assembled the resolved package. Telangana built the design cluster. Odisha moved decisively when the moment arrived. The rest of India offered incentives. The semiconductor map is the clearest available evidence that in India's federal industrial architecture, institutional execution at the state level determines which central policy ambitions become physical factories and which remain announcements.