India's semiconductor strategy is shifting from incentivising fabs to funding application-specific design and commercialisation. Critically examine with reference to the RDI Scheme.

The Research Development and Innovation (RDI) Scheme has a ₹1 lakh crore corpus and was approved on 1 July 2025. It offers long-term, low-cost capital for higher-TRL innovation [2]. Under it, TDB-DST is giving ₹130 crore to BigEndian Semiconductors for Project VeerAI, a camera-focused AI Vision SoC [1]. This shows the shift is real, but it adds to fab-building rather than replacing it.

Evidence of the shift: strengths

• Bridging the "valley of death": Product DLI support is capped at ₹15 crore per application [3]. VeerAI costs ₹260 crore, and RDI funds the costly TRL-5 → TRL-9 stage that leads to market [1].
• Application-specific focus: The first use is CCTV in India and the Global South. Defence, automotive and medical uses come later, so the chip has a ready market instead of competing with general-purpose chips [1].
• Fiscal prudence: The support is Optional Convertible Debt that covers about half the cost [1]. The Scheme prefers low- or nil-interest loans to grants, so public money can come back [2].
• Scale and continuity: TDB has approved 22 projects with ₹2,192 crore of RDIF support [4]. These build on the 24 DLI design projects, one of which is in video surveillance [3].

Limitations

• Design sovereignty, not supply-chain sovereignty: No foundry is named [1]. Domestic fabs and OSAT units under ISM are still being built, so a Bengaluru-designed chip will probably be manufactured abroad. That is why fab support is still needed.
• Institutional seams: DLI (MeitY) and RDI (DST) have no formal handoff rule between them. Startups face two appraisals, and nobody checks for double funding [1][3].
• Debt stress: Repayment may begin before the product earns. The interest rate and moratorium have not been made public [1].
• Crowding-out risk: Private industry's share of GERD rose to 51.8% in 2023-24 [5]. Without data on additionality, it is unclear whether public money is adding to private R&D or replacing it.
• Civil liberties: Cheaper AI-enabled CCTV must pass the Puttaswamy (2017) proportionality test and the DPDP Act, 2023.

In short, RDI fills the design-to-market gap that fab incentives left open, but it cannot stand in for manufacturing capacity. A way forward would be a clear DLI-to-RDI pipeline, public disclosure of milestones and co-investment for each project, and privacy-by-default conditions. Together these would turn India's design strength into the Atmanirbhar, rights-respecting technology leadership that SDG-9 calls for.