TDB-DST signs agreement with BigEndian Semiconductors for ₹130 crore RDI support to develop indigenous AI Vision SoC

1. At a Glance

• The Technology Development Board (TDB) under the Department of Science & Technology (DST) has agreed to give ₹130 crore from the Research Development and Innovation (RDI) Fund to BigEndian Semiconductors, Bengaluru. The money funds Project VeerAI, which will build a camera-focused AI Vision System-on-Chip (SoC). [1]
• The topic joins three examinable threads: semiconductor self-reliance, "technology sovereignty" in surveillance and critical infrastructure, and how the ₹1 lakh crore RDI Scheme puts money into private R&D. [1][2]
• It shows a policy shift: the government is backing application-specific chips rather than general-purpose ones, and funding later-stage commercialisation (TRL-5 → TRL-9), not only early research. [1]

2. Why in the News

• On 5 October 2026, TDB-DST signed the agreement with M/s BigEndian Semiconductors for ₹130 crore of RDI support. Total project cost is ₹260 crore. [1]
• The support comes as Optional Convertible Debt (OCD). [1]
• This is one of the projects TDB is funding in its role as a Second-Level Fund Manager (SLFM) of the RDI Scheme. [3]

3. Background & Evolution

• 1995: Parliament passes the Technology Development Board Act, 1995 (Act 44 of 1995). [5]
• 1996: TDB is set up under DST. Its mandate is to commercialise indigenous technology, or to adapt imported technology for wider domestic use. [5]
• Semicon India Programme / DLI Scheme (MeitY): built up a domestic chip-design (fabless) base. This is the earlier design-stage support that VeerAI-type projects grow out of. [4]
• 1 July 2025: the Union Cabinet approves the RDI Scheme with a ₹1 lakh crore corpus. Its aim is more private-sector R&D and financing for high-risk, high-impact innovation, including deep-tech startups. [2]
• 4 February 2026: TDB opens its RDI call for proposals. BIRAC follows on 13 February 2026. [3]
• March 2026: ₹500 crore is disbursed to TDB for investment in eligible technology entities. [3]
• 5 October 2026: the Project VeerAI agreement is signed. [1]

4. Core Static Facts

Item	Fact
Funding body	Technology Development Board (TDB), DST, Ministry of Science & Technology [1]
Statutory base of TDB	TDB Act, 1995; TDB set up in 1996 [5]
Beneficiary	M/s BigEndian Semiconductors, Bengaluru, Karnataka [1]
Project	Project VeerAI: camera-focused AI Vision SoC on a "sovereign technology platform" [1]
Support amount	₹130 crore (RDI Fund) [1]
Total project cost	₹260 crore (so the RDI share is about 50%) [1]
Instrument	Optional Convertible Debt (OCD) [1]
TRL progression	TRL-5 → TRL-9 [1]
First use	CCTV and surveillance, in India and the Global South [1]
Later sectors	Defence, automotive, industrial, medical [1]
Secretary, TDB	Sh. Rajesh Kumar Pathak [1]
RDI Scheme corpus	₹1 lakh crore; Cabinet approval on 1 July 2025 [2]
RDI structure	Two tiers: (1) a Special Purpose Fund (SPF) within ANRF, which holds the funds; (2) Second-Level Fund Managers [2]
SLFMs named	TDB and BIRAC [3]
TDB's RDI pipeline	22 projects approved; total project cost ₹4,744 crore; RDIF support ₹2,192 crore [3]
TDB's traditional instruments	Soft loans (up to 50% of project cost), equity (up to 25%), grants in exceptional cases [5]

Key terms (conceptual):

• SoC (System-on-Chip): one chip that combines processor, memory interfaces, accelerators (here, AI and vision) and connectivity.
• TRL (Technology Readiness Level): a 9-step maturity scale. TRL-5 means a component is validated in a relevant environment. TRL-9 means the system is proven in real operation.
• OCD (Optional Convertible Debt): a loan that the lender can choose to convert into equity. The public funder gets possible upside without taking an equity stake on day one.

5. Multi-Dimensional Analysis

Scientific / Technological

• Project VeerAI combines AI processing, smart features, security, privacy and connectivity on a single chip. [1]
• It is deliberately application-specific. The aim is faster deployment and commercialisation, not a contest with general-purpose chips. [1]
• The project builds on India's VLSI design talent. ChipIN's centralised EDA facility serves about 1 lakh engineers and students across 400 organisations. [1][4]

Strategic / Security

• The stated aim is to reduce dependence on externally developed semiconductor and technology supply chains. [1]
• Surveillance cameras sit inside critical infrastructure, so a trusted chip at device level helps with data security and technology sovereignty. [1]
• Targeting the Global South points to export ambitions and a possible "trusted-vendor" offer to developing countries. [1]

Economic

• The government covers about half of a ₹260 crore project, and the rest is private money. This is a co-investment model. [1]
• The RDI Scheme is meant to fix India's low private-sector share of R&D by offering patient, long-term capital. [2]
• A roadmap into several sectors (defence, automotive, medical) widens the market for Indian chips. [1]

Governance / Institutional

• The money moves in layers: ANRF (SPF) → TDB (SLFM) → company. This uses TDB's experience in appraising technology commercialisation. [2][3][5]
• Using convertible debt instead of grants protects public money and lets the state share in any success. [1]

Ethical

• AI chips built for CCTV surveillance raise questions about privacy and proportionality. The proposed Digital Personal Data Protection framework and the Puttaswamy (2017) privacy standard are relevant here. "Privacy features" in the hardware is a claim that will need independent checking.

6. Recent Developments (last 12–18 months)

• 1 July 2025: Cabinet approves the ₹1 lakh crore RDI Scheme. [2]
• 4 and 13 February 2026: TDB and BIRAC open their RDI calls for proposals. [3]
• March 2026: ₹500 crore released to TDB. [3]
• 2026: TDB approves 22 RDI projects worth ₹4,744 crore in total, with ₹2,192 crore of RDIF support. [3]
• Ongoing: the DLI Scheme has 24 chip-design projects, including one in video surveillance, with 16 tape-outs and 10 patents. [4]
• 5 October 2026: the Project VeerAI agreement is signed. [1]

7. Prelims Hooks

• Project VeerAI is funded by TDB under DST, not by MeitY or the India Semiconductor Mission. [1]
• The beneficiary, BigEndian Semiconductors, is based in Bengaluru, Karnataka. [1]
• RDI support is ₹130 crore against a total project cost of ₹260 crore. [1]
• The instrument is Optional Convertible Debt (OCD). [1]
• The project aims to move from TRL-5 to TRL-9. [1]
• The chip is a camera-focused AI Vision SoC, first for CCTV/surveillance. [1]
• The RDI Scheme has a corpus of ₹1 lakh crore and was approved by Cabinet on 1 July 2025. [2]
• RDI first tier: a Special Purpose Fund within ANRF, which holds the funds. [2]
• RDI Second-Level Fund Managers include TDB and BIRAC. [3]
• TDB was set up in 1996 under the TDB Act, 1995. [5]
• TDB is described as the first organisation of its kind within the Government whose sole aim is to commercialise indigenous research. [5]
• Under the DLI Scheme, Product DLI reimburses up to 50% of eligible expenditure, capped at ₹15 crore per application. [4]
• Deployment-linked incentive under DLI: 6% to 4% of net sales turnover for five years. [4]
• The DLI Scheme is run by MeitY under the Semicon India Programme. [4]

8. Why the Old ₹15 Crore Chip Grant Could Not Carry a Chip to Market

• The earlier scheme was built for designing a chip, not for selling one
• Under MeitY's DLI Scheme, the Product DLI pays back at most 50% of costs, and never more than ₹15 crore per application [4].
• Project VeerAI alone costs ₹260 crore [1]. That is about 17 times the DLI ceiling.

• So a fabless startup (a company that designs chips but owns no factory) could get its first design made with DLI money. It still had nothing to pay for the much costlier work of turning that design into a product.

• This gap has a name: the "valley of death"

• TRL-5 to TRL-9 is the stage where a design that works in a lab has to work in thousands of real cameras [1].
• Research grants usually stop before this stage. Venture investors often wait until after it.

• The RDI Fund is aimed at exactly this middle stage [1][2]. That is why the deal matters more than its size suggests.

• The two schemes now sit one after the other, under two different ministries

• DLI (MeitY) has 24 chip-design projects, and one of them is already in video surveillance [4].
• RDI money for VeerAI comes from DST, through TDB [1].
• No public rule says how a DLI project moves on to RDI money. A startup has to apply to two different appraisers, and nobody checks whether the same costs are funded twice.

9. The Loan Protects Public Money but Puts Pressure on a Young Company

• Why the RDI Fund gives loans and not grants
• The Scheme mainly allows long-term loans at low or nil interest, plus equity only in some startup cases [2].

• If the project fails, part of the money can still come back. Grant money never does.

• The weak point: a loan has to be repaid, even before the product earns

• A chip firm earns little until the chip is proven and selling (TRL-9). That is the very stage this money is meant to reach [1].
• If repayments start before sales begin, the company may have to cut R&D, or raise costly private money, just to pay the loan.

• The press release does not give the interest rate, the moratorium (the period before repayment starts) or the conversion terms [1]. Without these, nobody can judge whether the deal is gentle or harsh on the company.

• Converting the loan into shares (OCD) makes the state an owner of a surveillance-chip company

• With Optional Convertible Debt, TDB decides whether to turn the loan into shares [1].
• If it converts, a government body becomes a shareholder in a firm whose first market is CCTV, including sales abroad [1].

• That brings a conflict: the same state that buys and regulates surveillance cameras would also gain if more of them are sold.

• TDB has two roles in the same deal

• TDB judges which proposals are good, and it also puts in the money [3][5].
• TDB has approved 22 projects worth ₹2,192 crore of RDI support, while it has received only ₹500 crore so far [3].
• Its approvals are therefore already far ahead of the cash it has. How fast the Special Purpose Fund in ANRF releases more money will decide whether these approvals actually get paid out [2][3].

10. If Private R&D Is Already Rising, What Does This Fund Add?

• The strongest objection
• Private industry's share of GERD (Gross Expenditure on R&D, India's total R&D spending) rose from 45.5% in 2021-22 to 51.8% in 2023-24 [7].
• Private firms now spend more on R&D than the government does [7]. A critic can ask: why put cheap public money into a private company that might have raised it anyway?

• Economists call this the risk of displacing private money. Public money simply replaces money investors would have given, so no new research happens.

• What the critic gets right

• The release says nothing about whether BigEndian tried to raise this money privately and failed [1].

• Any company would happily take a 50% public share at low interest [1][2]. Being eager to take the money does not prove the money was needed.

• Why the objection is not the full story

• A rising share of private R&D says nothing about which sectors get that money. A higher national figure does not mean chip design gets more of it.

• Chip development is costly, slow and binary: a design either works after manufacture or it does not. That is the kind of risk the RDI Scheme says it is meant to carry [2].

• What would settle the question

• TDB should publish, project by project, how much private money came in alongside the RDI money, and whether each project hit its TRL milestones on time.
• Without this, the 22-project pipeline [3] shows only that money was spent, not that it created research that would not otherwise have happened.

11. A Sovereign Chip, Made Where?

• Designing a chip in India is not the same as owning the whole supply chain
• The project's stated aim is to cut dependence on supply chains controlled outside India [1].
• But the release names no foundry (the factory that actually manufactures the chip) [1].

• India's own fab and ATMP/OSAT units (where chips are assembled, tested and packaged) are still being built under the India Semiconductor Mission. For now, a chip designed in Bengaluru will probably be manufactured abroad.

• Why this matters for the security argument

• The case for a "trusted chip at device level" is that India controls what is inside it [1].
• If the chip is made in a foreign factory, the design is Indian but the making of it is not. The "sovereignty" covers the design and the software, not the full chain.
• For a Mains answer, write "design sovereignty". Do not claim full self-reliance.

12. A Privacy Feature in the Chip Does Not Control Who Watches the Footage

• The chip makes surveillance cheaper and smarter. That is the point, and also the risk
• An AI Vision SoC lets a camera recognise people, objects or events by itself, without sending all footage to a central server [1].

• Cheaper smart cameras mean more of them. Each one can now analyse what it sees, not just record it.

• Hardware cannot decide what is lawful

• "Privacy features" on a chip can encrypt footage or blur faces [1]. But the person who buys the camera decides whether to switch those features on.

• The Puttaswamy (2017) test asks whether surveillance is backed by law, needed, and proportionate (no more intrusive than the aim requires). A chip cannot answer any of these questions. Only rules on how the camera is used can.

• Selling to the Global South carries this question abroad

• The first export market is CCTV in developing countries [1]. Many of these countries have weak data-protection laws.

• If India sells itself as a "trusted vendor", that trust will be judged by how the chips are used, not only by where they were designed.

• What should follow

• TDB, as funder and possible shareholder, could make public funding conditional on privacy features being switched on by default, with independent testing of that claim.
• The DLI Scheme already pays incentives only after products are actually deployed [4]. A similar rule could link the later RDI payments to proven privacy safeguards, not only to TRL milestones.

13. Anchors for Answers

• Data: Private industry's share of India's R&D spending (GERD) reached 51.8% in 2023-24, up from 45.5% in 2021-22 [7]
• Data: VeerAI's ₹260 crore total cost against the Product DLI cap of ₹15 crore per application shows the funding gap between design and market [1][4]
• Data: TDB has approved ₹2,192 crore of RDI support but received only ₹500 crore so far, so approvals are well ahead of the cash [3]
• Report/Committee: DST, R&D Statistics at a Glance 2025-26 (compiled by NSTMIS) [7]
• Law/Case: K.S. Puttaswamy v. Union of India (2017), the legality-necessity-proportionality test for surveillance; Digital Personal Data Protection Act, 2023; Technology Development Board Act, 1995 [5]
• Scheme: Design Linked Incentive Scheme (MeitY). It funds the design stage and links payments to deployment; RDI funds the later TRL-5 → TRL-9 stage [4][1]

14. Mains Relevance

• GS-III: Science & Technology (indigenisation of technology, developing new technology); Indian Economy (industrial policy, investment models); Internal Security (security of critical infrastructure, role of technology).
• GS-II: Government policies and interventions (design of the RDI Scheme, ANRF).
• GS-IV: Ethics of surveillance technology.

Question stems:

1. India's semiconductor strategy is shifting from incentivising fabs to funding application-specific design and commercialisation. Critically examine with reference to the RDI Scheme. (250 words)
2. How does the multi-tier structure of the Research Development and Innovation Scheme try to correct the low private-sector share in India's R&D spending? Discuss its limitations. (150 words)
3. Indigenous AI chips for surveillance strengthen technology sovereignty but raise civil-liberty concerns. Discuss. (150 words)

15. Related Topics to Study Next

• Anusandhan National Research Foundation (ANRF): it holds the RDI Special Purpose Fund.
• India Semiconductor Mission / Semicon India Programme: support for fabs, ATMP/OSAT and display units, alongside design.
• Design Linked Incentive (DLI) Scheme: the earlier, design-stage support for fabless startups. [4]
• BIRAC (DBT): the other RDI Second-Level Fund Manager, for biotech. [3]
• Deep Tech Fund of Funds: a separate deep-tech financing route, often confused with RDI.
• IndiaAI Mission: compute and model-building, which complements AI hardware.
• Digital Personal Data Protection Act, 2023: the privacy rules that apply to AI-enabled surveillance.
• National Technology Day / National Technology Awards (TDB): TDB's recognition role. [5]

16. NCERT Concepts to Revise

• Design-linked incentive (Economy): The earlier chip-design support scheme, with its ₹15 crore cap, explains the funding gap VeerAI's RDI money fills.
• India Semiconductor Mission (Geography): India's main chip programme for fabs and packaging units; the note contrasts it with this DST-funded design project.
• Deep-tech startup (Economy): BigEndian is a deep-tech firm needing long, costly R&D, which is what the RDI Scheme targets.
• Doctrine of proportionality (Polity): The Puttaswamy test for judging whether AI-enabled CCTV surveillance is lawful, necessary and no more intrusive than needed.
• Sovereignty (Polity): The project is justified as technology sovereignty, so you should know what sovereignty means and where design-only control falls short.
• Digital Personal Data Protection Act 2023 (Polity): The privacy law that governs how AI-enabled cameras may collect and process people's personal data.

17. Common Errors / Trap Areas

• Wrong ministry: semiconductors are usually linked to MeitY/ISM, but this funding comes from DST's TDB through the RDI Fund. [1]
• RDI Fund ≠ ANRF grants: ANRF holds the SPF, but the money reaches companies through SLFMs like TDB and BIRAC, mostly as loans or convertible instruments rather than grants. [2][3]
• ₹130 crore vs ₹260 crore: ₹130 crore is the RDI support; ₹260 crore is the total project cost. [1]
• TRL direction: the project goes TRL-5 → TRL-9 (towards commercial deployment), not TRL-1 → 5. [1]
• OCD is not equity: it starts as debt and becomes equity only if the lender chooses to convert.