Wrong Diagnosis, Right Fund: What India's DeepTech Debate Is Missing

Part I: Beyond Categories, Towards Capabilities

On July 2, 2026, at the launch of the AI Council of India in Mumbai, senior economist and policy adviser Neelkanth Mishra said that India's ₹1 lakh crore Research, Development and Innovation (RDI) Fund has struggled to identify enough deserving DeepTech startups to absorb capital at the scale originally envisaged. He says that India's DeepTech ecosystem remains immature, and the venture-capital community has yet to develop the expertise needed to evaluate and finance frontier technologies.

His observation is correct on its own narrow terms — and that is precisely the problem. 

As a description of the startup-financing pipeline, this is accurate. As a diagnosis of India's innovation capacity, it mistakes a symptom for the disease. A shortage of investable DeepTech startups is not evidence that India lacks deserving projects, or even deserving institutions. It is evidence that we have let one financing category — startups — stand in for the entire innovation ecosystem, and then judged that ecosystem's health by how well the category is performing. A shortage of applicants is not the same as a shortage of institutions.

To be fair, Mishra's own account is not confined to venture capital alone. He has pointed to the need for stronger, merit-based research ecosystems at institutions like the IITs and IISc, and linked this to reverse brain drain, stronger domestic IP creation, and globally competitive products in strategic sectors. This is a wider frame than pure VC capacity, and it deserves acknowledgment.

But even this wider frame stays inside a single model. Research institutions matter, in his account, insofar as they make venture capital more confident backing startups. The pipeline gets a longer front end — better talent, better IP, better-prepared founders. It does not become a different shape. Nowhere in this account do we find translation infrastructure that converts research into manufacturable technology, legacy industry and GCCs as capability generators in their own right rather than startup feeder systems, or the coordinating institutions that allow capability to diffuse across an industry rather than accumulate inside one company. Widening the pipeline's inputs is not the same as questioning whether or not a pipeline is the right architecture at all.

This matters because Mishra's framing has quietly become the framing of the policy conversation itself. Discussions of the RDI Fund increasingly assume it exists primarily to finance DeepTech startups — and, by that measure, are increasingly asking whether it is failing.

It is not failing. It is being asked the wrong question.

India's venture-capital ecosystem has historically evolved around consumer internet businesses, fintech, e-commerce, and software services. Investing in semiconductors, advanced materials, industrial biotechnology, robotics, energy technologies, or advanced manufacturing requires very different technical understanding, longer investment horizons, and greater tolerance for technological uncertainty. That much of Mishra's underlying point is fair — building such an investment ecosystem will take time.

But conceding that point should not mean conceding the frame. If we allow startup-financing capacity to define whether the RDI Fund is working, we will spend the next several years trying to manufacture more fundable startups, while the actual constraint — a missing institutional middle between research and industrial deployment — remains untouched.

Have we, in effect, begun treating the ₹1 lakh crore RDI Fund as though it were primarily a DeepTech startup financing mechanism?

That framing is understandable. After all, the RDI Fund has been designed to crowd in private investment into industry-led research and innovation, particularly at later stages of technological development where promising ideas require patient capital to move towards commercial deployment.

Yet allowing the financing architecture to define the policy conversation carries its own risks. Because then the discussion gradually shifts from asking what capabilities India needs to build to asking who should receive the money. Once that happens, startups naturally become the centre of attention. The larger innovation ecosystem begins to fade into the background.

India's technological future, however, will not be built by startups alone. Nor, for that matter, is "DeepTech" itself a coherent industrial sector.

Both assumptions deserve closer examination.

DeepTech Is Not an Industry

One reason the current discussion has become unnecessarily narrow is that the term DeepTech itself has gradually acquired a misleading meaning. It is increasingly spoken of as though it represents a distinct industrial sector.

It does not.

In many ways, it resembles the term MSME. Nobody describes MSMEs as an industry. They include textile manufacturers, engineering workshops, machine-tool companies, restaurants, logistics firms, pharmaceutical manufacturers, software companies, food processors, and thousands of other enterprises. MSME is an administrative classification based on enterprise characteristics rather than industrial activity.

DeepTech functions similarly. It is not an economic sector but a broad investment category encompassing businesses whose competitive advantage depends upon significant scientific and engineering capability.

Within this umbrella lie industries as diverse as:

semiconductors,

battery technologies,

artificial intelligence,

industrial robotics,

biotechnology,

aerospace,

advanced materials,

industrial electronics,

photonics,

medical technologies,

quantum technologies,

advanced manufacturing,

and multiple emerging engineering disciplines.

These industries share one broad characteristic: they are technologically intensive. Beyond that, they differ enormously.

A semiconductor fabrication ecosystem bears little resemblance to a biotechnology manufacturing ecosystem. Battery technologies require institutions very different from those needed for quantum computing. Industrial AI follows commercialisation pathways unlike those of advanced materials or medical devices.

Each demands its own combination of:

scientific expertise,

engineering capabilities,

manufacturing systems,

testing infrastructure,

certification frameworks,

regulatory institutions,

industrial customers,

capital intensity,

and supply-chain ecosystems.

Grouping them together under one label is administratively convenient. But it can become analytically misleading. A single "DeepTech policy" can only go so far because DeepTech itself is not a single technological or industrial domain.

DeepTech Is Bigger Than Startups

Once we recognise that DeepTech is not an industry, another assumption also begins to unravel.

Public discussions increasingly treat DeepTech as though it were synonymous with startups. This is understandable. Over the past decade, startups have become the most visible face of technological entrepreneurship in India. Much of the discussion surrounding frontier technologies therefore assumes that India's technological future will primarily emerge through venture-backed firms.

But this captures only one part of a much larger ecosystem.

India's DeepTech landscape already extends well beyond startups. Large industrial companies are investing in batteries, advanced materials, industrial AI, electronics manufacturing, green hydrogen, precision engineering, and next-generation manufacturing systems. Public sector enterprises continue expanding into strategically important technologies related to energy systems, defence manufacturing, transportation, power electronics, and advanced engineering. Universities and research institutions are steadily strengthening India's scientific base. Research laboratories continue contributing to strategic technologies across multiple domains. And increasingly, Global Capability Centres are moving beyond traditional support functions into advanced engineering, product development, and industrial research.

Each of these institutions contributes differently to technological capability. Startups excel at experimentation and rapid exploration. Large industrial firms possess manufacturing experience, deployment capability, supplier relationships, and long-term capital. PSUs often undertake strategic projects whose commercial returns may take years to materialise. Universities generate foundational scientific knowledge and specialised talent. Research institutions accumulate long-term technical expertise. GCCs increasingly connect global technological networks with Indian engineering capabilities.

None of these institutions alone constitutes India's DeepTech ecosystem. Rather, India's technological strength emerges from the interactions among them.

Recognising this distinction fundamentally changes how we should think about the RDI Fund. Instead of asking:

"Which organisational category should receive public support?"

India should begin by asking a different question:

Which capability-building functions require strengthening?

That shift — from organisational categories to national capabilities — opens up a much broader conversation. It encourages us to think not only about startups, but also about legacy industry, universities, research institutions, Global Capability Centres, public-sector enterprises, and the shared institutions that connect them into one coherent innovation ecosystem.

It is this broader capability architecture, rather than any single organisational form, that will ultimately determine whether India succeeds in translating scientific knowledge into long-term technological and industrial strength.

Part II: Beyond Startups — The Other Builders of Technological Capability

If India's DeepTech ecosystem extends well beyond startups, then the obvious question follows.

Who are the other institutions contributing to India's technological capabilities?

Two deserve particular attention because both are frequently misunderstood.

The first is the Global Capability Centre (GCC).

The second is the legacy industrial enterprise.

Both have become increasingly important to India's innovation landscape. Yet public discussions often fail to appreciate the roles they perform.

GCCs Are Not an Industry Either

The same conceptual mistake that is often made about DeepTech is increasingly being made about Global Capability Centres.

GCCs are frequently discussed as though they constitute an independent industry.

They do not.

Like DeepTech, GCCs represent an organisational category that cuts across numerous industries. India today hosts GCCs belonging to:

aerospace companies,

automotive manufacturers,

industrial engineering firms,

pharmaceutical companies,

chemical producers,

financial institutions,

healthcare companies,

electronics manufacturers,

energy companies,

and many other sectors.

Nor do they perform a single function. Some continue focusing primarily on software engineering, finance, procurement, customer support, or enterprise operations. Others undertake:

product engineering,

artificial intelligence research,

semiconductor design,

materials science,

industrial process engineering,

advanced manufacturing research,

and frontier technology development.

Treating all GCCs alike therefore obscures the remarkable diversity that now exists within India's GCC ecosystem. More importantly, it risks overlooking a growing source of technological capability.

A recent example illustrates this shift. Late last year, Honeywell licensed its ethanol-to-jet-fuel (ETJ) technology to an Indian company. Viewed superficially, this appeared to be another routine technology-licensing agreement between a multinational corporation and an Indian industrial partner.

The more interesting detail lay elsewhere. The technology itself had been developed at Honeywell's Global Capability Centre in Gurugram.

This is significant for two reasons. First, it demonstrates that certain GCCs are no longer simply supporting global operations. They are generating original engineering and industrial technologies. Second, it shows how technological capability developed in India can subsequently find industrial deployment within India.

This does not mean that every GCC should automatically qualify for public support. Far from it. Many GCCs continue performing functions that fall well outside the intended scope of the RDI Fund.

The point is different. Organisational identity alone should neither qualify nor disqualify a project. The relevant question should always remain:

Does the project strengthen India's long-term technological capabilities?

If the answer is yes, then the fact that the research originated within a GCC becomes secondary. The capability being created matters more than the institutional label attached to it.

Legacy Industry Is Part of the DeepTech Story

Large industrial enterprises are often viewed through an entirely different lens. Public commentary increasingly compares Indian conglomerates with Chinese technology champions and concludes that Indian companies simply need to invest more aggressively in frontier research.

Greater industrial R&D investment is certainly desirable. But such comparisons often overlook an important difference.

Large industrial companies do not innovate under the same conditions — or for the same purposes — as startups. Startups are designed to explore uncertainty. They pursue new ideas rapidly, experiment with emerging technologies, and often operate without established manufacturing systems or existing customer commitments.

Industrial enterprises face a different challenge. They must integrate innovation into:

large manufacturing systems,

complex supply chains,

regulatory environments,

operational reliability,

long investment cycles,

and commercial deployment at industrial scale.

Their innovation pathways are therefore necessarily different.

Moreover, the technologies they develop frequently generate benefits extending well beyond the originating company. Advances in:

battery chemistry,

industrial process engineering,

advanced metallurgy,

power electronics,

hydrogen systems,

industrial automation,

or precision manufacturing...

...rarely remain confined within one corporate boundary. They influence suppliers. They shape technical standards. They diffuse into universities and research collaborations. They strengthen manufacturing ecosystems. They create specialised engineering talent. And over time, they improve the technological capabilities of entire industrial sectors.

Economists describe these as knowledge spillovers. From a public-policy perspective, they provide an important justification for co-funding strategic industrial research.

Public support is not warranted because large companies lack financial resources. It is warranted because society captures far more value from successful frontier research than any single company can fully appropriate.

The objective is therefore not corporate subsidy. It is capability accumulation.

This distinction is important because it broadens the conversation beyond startups without diminishing their importance. Startups remain indispensable. But they are not the only institutions capable of generating frontier technologies. Nor are they always the best placed to undertake long-duration, capital-intensive industrial research requiring extensive manufacturing expertise.

India's technological future will almost certainly depend upon both. Entrepreneurial experimentation will continue producing new ideas. Large industrial enterprises will increasingly become responsible for scaling, integrating, manufacturing, and continuously improving many of those technologies over long periods.

Recognising this complementary relationship has important implications for the RDI Fund. If the objective is simply to finance startups, the current debate is sufficient. But if the objective is to strengthen India's long-term technological capabilities, then startups represent only one part of a much larger institutional landscape.

The challenge is not simply to finance individual innovators. It is to strengthen the broader ecosystem through which discoveries become technologies, technologies become products, and products become enduring industrial capabilities.

That broader perspective naturally leads to the next question.

Even if India possesses talented researchers, ambitious startups, capable industrial firms, increasingly sophisticated GCCs, and growing research institutions, why does the country still struggle to convert many promising technologies into globally competitive industries?

The answer lies not in the absence of scientific talent or entrepreneurial ambition. It lies in a set of institutions that India has repeatedly underbuilt.

It lies in what may be called the missing middle.

Part III: The Missing Middle

By this point, the discussion has established a broader proposition.

India's technological capabilities do not reside in startups alone. They are distributed across universities, research institutions, Global Capability Centres, legacy industrial companies, public-sector enterprises, and an expanding ecosystem of scientific and engineering organisations.

Yet a fundamental question remains. If India possesses growing scientific talent, increasing industrial ambition, expanding corporate R&D, and an increasingly diverse innovation ecosystem, why does it still struggle to convert many promising technologies into globally competitive industries?

The answer, I believe, lies in a layer of the innovation system that receives remarkably little public attention.

India has steadily strengthened its ability to generate scientific knowledge, engineering talent, entrepreneurial ambition, and increasingly sophisticated industrial research. What it repeatedly underbuilds are the institutions that connect one stage of capability to the next.

This is India's missing middle.

Across sectors, the pattern is strikingly similar. The country often succeeds in creating:

scientific discoveries,

engineering talent,

research laboratories,

startup activity,

industrial ambition,

and even manufacturing capacity.

But it struggles to build the translational infrastructure that systematically converts these into commercially deployable technologies.

These missing institutions include:

pilot plants,

process engineering centres,

testing and validation facilities,

certification ecosystems,

manufacturing demonstrators,

open-access prototyping facilities,

and industrial-scale experimentation platforms.

They are less visible than research laboratories and less glamorous than startup funding rounds. Yet they often determine whether an invention becomes an industry.

Why Translation Matters

India's semiconductor ecosystem offers perhaps the clearest illustration.

The Chips to Startup (C2S) programme dramatically expanded India's chip-design capabilities. Thousands of engineers received specialised training. Universities gained access to sophisticated electronic design automation tools. Hundreds of custom chips were designed.

The programme succeeded in creating design capability. Yet for several years, one uncomfortable question remained unanswered.

Where would these chips actually be fabricated?

The problem was never simply a shortage of engineers or ideas. The problem was the absence of an accessible fabrication bridge between design and manufactured silicon.

The upgradation of the Semiconductor Laboratory (SCL) in Mohali late last year therefore represented far more than another semiconductor investment. It created a translation institution. For perhaps the first time, Indian researchers, startups, and designers could realistically imagine moving from chip design to fabricated prototypes without depending entirely upon overseas facilities.

The lesson extends far beyond semiconductors.

Suppose an Indian university develops a breakthrough battery chemistry. Or a GCC engineers a novel sustainable aviation fuel process. Or a large industrial company discovers a superior hydrogen-production catalyst. Or a startup develops an advanced industrial robot.

Scientific success immediately generates another set of questions.

Who validates the manufacturing process? Who produces pilot-scale batches? Who certifies the technology? Who stress-tests the system under industrial operating conditions? Who demonstrates that it can be manufactured reliably, economically, and repeatedly?

Without institutions capable of answering these questions, research frequently produces papers, patents, prototypes, and promising announcements. It does not necessarily produce industries.

The constraint is no longer knowledge. It is translation.

The RDI Fund Already Recognises This Challenge

Interestingly, the design of the RDI Fund implicitly acknowledges this reality.

Unlike conventional research grants, the scheme focuses on projects that have already progressed beyond basic scientific discovery. Its emphasis on Technology Readiness Level (TRL) 4 and above reflects a recognition that India's challenge increasingly lies in moving promising technologies towards commercial deployment.

This is precisely where the missing middle begins.

The opportunity, therefore, is not to redefine the purpose of the RDI Fund. Its existing mandate is already aligned with one of India's most important technological bottlenecks. The opportunity is to broaden how that mandate is interpreted.

At present, discussions naturally gravitate towards financing individual projects. But translation depends upon more than projects alone. It also depends upon shared institutional infrastructure.

A pilot plant does not belong to one technology. A validation centre rarely serves only one company. A certification ecosystem benefits an entire industry. An open-access manufacturing demonstrator lowers barriers for universities, startups, industrial firms, GCCs, and public-sector enterprises alike.

These are ecosystem assets. Individual firms often cannot justify building them because the benefits extend well beyond their own commercial interests. Yet without them, even well-funded research programmes struggle to move beyond the laboratory.

The missing middle, therefore, is not merely an absence of infrastructure. It is an absence of institutions capable of systematically translating scientific capability into industrial capability.

Recognising this distinction changes how the RDI Fund itself should be understood. Rather than viewing it simply as a financing mechanism for promising technologies, it can also be seen as an opportunity to strengthen the institutional bridges through which those technologies become manufacturable, deployable, and commercially scalable.

That shift — from financing projects alone to strengthening the ecosystem within which projects mature — may ultimately prove to be one of the Fund's most important contributions.

The question, therefore, is no longer whether India should invest in research. Nor is it whether startups deserve support. The more important question becomes:

How should India organise public investment so that knowledge, wherever it is created, can reliably become long-term national capability?

Part IV: From Project Financing to National Capability

The discussion thus far leads to a broader proposition.

The ₹1 lakh crore RDI Fund should not be judged merely by how many projects it finances. Its larger significance lies in the opportunity it presents to strengthen India's long-term capability architecture.

Importantly, this does not require redefining the purpose of the scheme. The RDI Fund has already been designed to crowd in private investment into research, development and innovation, particularly for technologies that have progressed beyond basic scientific discovery and require patient capital to move towards commercial deployment.

The opportunity lies elsewhere. It lies in broadening how eligible capability-building investments are understood. Instead of viewing the Fund primarily as a mechanism for financing individual organisations, India can begin viewing it as an instrument for strengthening the ecosystem through which technological capability accumulates.

This requires moving beyond organisational categories and focusing instead on capability-building functions.

Three Windows for Building National Capability

One possible way of interpreting the RDI Fund is through three complementary capability-building windows.

1. Capability Creation

The first window supports the creation of new technological knowledge. Eligible projects may involve:

startups,

universities,

research institutions,

Global Capability Centres,

legacy industrial companies,

public-sector enterprises,

and collaborative research consortia.

The relevant question is not who conducts the research. It is whether the proposed work advances India's technological capabilities in strategically important domains. Projects led by a startup, an industrial company, or a GCC should therefore be evaluated on their technological contribution rather than their organisational identity.

2. Capability Translation

The second window addresses the institutional gap that has repeatedly appeared throughout this discussion. It supports the conversion of promising research into manufacturable technologies. This includes investments in:

pilot plants,

process engineering,

testing facilities,

validation centres,

certification systems,

manufacturing demonstrators,

hardware-in-the-loop facilities,

and other shared translational infrastructure.

These assets perform a unique function. They reduce the cost and uncertainty of moving technologies from laboratories into commercial production. Unlike conventional industrial assets, they often serve entire ecosystems rather than individual firms. This is precisely why public participation becomes economically justified.

The objective is not to replace private investment. It is to strengthen the shared institutional infrastructure upon which private investment ultimately depends.

3. Capability Diffusion

The third window ensures that technological capability spreads throughout the wider economy. Industrial competitiveness is not determined simply by who invents new technologies. It also depends upon how effectively those technologies become embedded across firms, suppliers, workers, and production systems.

This window therefore supports activities such as:

technology-transfer partnerships,

collaborative demonstration projects,

supplier capability development,

workforce skilling,

industrial adoption,

cluster-based innovation,

standards implementation,

and large-scale commercial deployment.

Successful innovation is ultimately measured not by isolated breakthroughs but by widespread capability accumulation.

Capability Coordination: The Institutional Glue

These three capability-building windows cannot operate independently. Each depends upon an institutional function that has historically received far less attention than research or financing.

That function is coordination.

Coordination is not another stage occurring after research or before commercialisation. It is the institutional glue that enables every stage of capability development.

Before research begins, someone must identify strategic technological priorities, formulate research roadmaps, assemble collaborative consortia, and align public and private investment. During translation, someone must govern shared facilities, establish technical standards, coordinate industrial participation, and ensure open access to expensive infrastructure. During diffusion, someone must organise supplier development, workforce training, technology transfer, industrial clusters, and standards adoption.

These are fundamentally coordination challenges. They cannot be solved by individual firms acting independently.

This is where industry associations deserve far greater attention than they usually receive. Traditionally viewed as representative organisations, they can increasingly become institutional orchestrators of industrial capability.

Working alongside government, universities, research institutions, startups, GCCs, public-sector enterprises, and industrial firms, they can:

identify common technological bottlenecks,

formulate sector-wide research priorities,

coordinate collaborative R&D programmes,

govern shared translational infrastructure,

strengthen supplier ecosystems,

develop technical standards,

facilitate international technology partnerships,

and accelerate technology diffusion across industries.

Many of the shared facilities discussed earlier — from battery validation centres and advanced materials pilot plants to robotics testing facilities and industrial electronics demonstrators — would likely function best under collaborative governance involving industry associations, research institutions, and government rather than being owned exclusively by any one organisation.

Supporting such coordinating institutions is therefore not an administrative expense. It is itself an investment in national capability.

Organising Around Functions, Not Categories

Seen through this lens, many contemporary innovation debates begin to look unnecessarily constrained.

DeepTech is not an industry. Startups are not an industry. Global Capability Centres are not an industry. MSMEs are not an industry. These are organisational, administrative, or investment categories that cut across multiple sectors. Yet policy discussions frequently treat them as though they constitute separate domains deserving independent funding frameworks.

A capability-based approach offers a more coherent alternative. Instead of asking, "Should startups receive more support?" the question becomes, "Which capability-building function does this project perform?" Instead of asking, "Should GCCs be eligible?" the question becomes, "Does this project strengthen India's long-term technological capabilities?"

The same principle applies to industrial companies, universities, public-sector enterprises, research laboratories, and industry associations. Projects should be assessed by the capability they create, translate, or diffuse — not by the organisational category of the applicant.

The result is a far more flexible and future-ready framework. A startup developing next-generation robotics could qualify. A university leading advanced semiconductor research could qualify. A conglomerate investing in frontier battery chemistry could qualify. A PSU developing advanced grid technologies could qualify. A GCC developing industrial technologies in partnership with Indian manufacturing could qualify. An industry association establishing a shared pilot facility or coordinating a sector-wide technology programme could equally qualify.

The common criterion is contribution to national capability. Not institutional identity.

Conclusion: Beyond Innovation, Towards Capability Accumulation

Much of the discussion surrounding the RDI Fund focuses on innovation. Innovation is undoubtedly important. But innovation alone is episodic. A breakthrough discovery, a successful startup, or a patented technology represents an important achievement. Yet countries become technological powers not through isolated breakthroughs, but through the steady accumulation of capabilities over decades.

Capability accumulation asks a different question. Is a country becoming progressively better at discovering, engineering, translating, manufacturing, coordinating, deploying, improving, and diffusing increasingly sophisticated technologies throughout its economy? That requires institutions as much as inventions, infrastructure as much as investment, coordination as much as competition — and, above all, continuity.

Neelkanth Mishra is right that India currently lacks enough investable DeepTech startups. But he is treating that shortage as the diagnosis, rather than as a symptom of a much larger design flaw: a fund built to strengthen national capability has been evaluated as though its only job were to feed the venture pipeline. Judged on that basis, of course it looks like the fund is falling short. It was never actually asked to do the thing it is being measured against.

The ₹1 lakh crore RDI Fund presents India with a rare opportunity — not merely to finance more DeepTech projects, but to strengthen the institutional architecture through which knowledge becomes engineering capability, engineering capability becomes industrial production, and industrial production becomes enduring national technological strength.

The Fund's greatest contribution ultimately may not lie in financing DeepTech projects, but may lie in systematically catalysing national capabilities.