Deeptech's Hardware Problem: ‘Make in India’ Can't Survive Without Scale

India’s push toward sovereign AI and hardware manufacturing is at an inflection point. Policymakers are investing billions, startups are building indigenous stacks, and there’s a narrative around self-reliance. 

But beneath the ambition lies a harder question: can India’s hardware ecosystem sustain itself if its own customers fail to scale?

At the heart of this tension is a structural dependency and demand-side fragility that few openly acknowledge. As Raghu Venkatesh, Co-founder and CBO at ANSCER Robotics, puts it, hardware manufacturers cannot scale in isolation. If clients don’t scale, sovereign AI hardware manufacturers get no business.

The Demand Problem

India’s deeptech hardware ecosystem, spanning drones, robotics, and AI infrastructure, has largely been built on the assumption that domestic demand will follow innovation. But that assumption is increasingly being challenged.

The country’s deeptech market is expected to reach $30 billion by 2030, per IBEF, but a large part of the boom can be attributed to advances in defence innovation, government indigenisation, and infrastructure upgradation. However, domestic procurement remains weak.

“The biggest challenge, at least in India, is customer adoption,” Manish Gupta, General Partner at GrowX, tells AIM. “Indian corporates don’t even spend enough time, effort, or money to do pilots with these robotic companies.”

This lack of early adoption creates a cascading effect. Without pilots, there is no validation. Without validation, there is no scale. And without scale, hardware manufacturers, especially those attempting contract manufacturing, are left without predictable revenue streams.

Even when startups build full-stack hardware solutions, selling them remains difficult. “Forget about scale. We don’t buy the first products,” Gupta vents.

Meanwhile, the idea that Indian hardware companies can sustain themselves by manufacturing for other startups or enterprises isn’t as sound either. In theory, contract manufacturing offers a scalable, capital-efficient model. In practice, it is deeply tied to the fortunes of client companies. If those clients fail to scale, demand collapses.

Manufacturing capacity without downstream demand is uneconomical. Building hardware in India is still more expensive than importing from established global supply chains. “If somebody is trying to save cost, they will simply go and buy it from China,” Gupta notes.

This creates a vicious cycle: low demand prevents scale, lack of scale keeps costs high, and high costs further suppress demand.

Sovereignty vs Reality

The broader narrative of sovereign AI adds another layer of complexity. India’s policy ecosystem—through initiatives like the India Semiconductor Mission and the IndiaAI Mission—aims to reduce dependence on foreign technology. But experts caution against overestimating current capabilities.

Bharath Reddy, a research fellow at the Takshashila Institution, argues that India’s progress is foundational, not transformational. “These are not enough for strategic autonomy or supply chain resilience,” he says, referring to existing semiconductor and AI initiatives.

India has made strides, including creating a pool of over 38,000 GPUs and investing $10 billion into semiconductor infrastructure. But these efforts are building “base-level capabilities,” not full-stack sovereignty.

More importantly, Reddy challenges the very premise of complete self-reliance. “AI or compute sovereignty is aspirational, not realistic at current investment levels and structural constraints,” he says. 

Currently, India remains highly import-dependent for AI and hardware, with semiconductor imports totalling $20.6 billion in FY24, up 18.5% year-on-year. Over 90% of advanced chips, GPUs, sensors, and fabrication equipment are sourced globally, with limited domestic fabs and a reliance on 40+ countries for critical electronics supply chains.

Beyond Manufacturing

Instead, Reddy believes India’s strategy should focus on selective capability-building while remaining deeply integrated with global supply chains. He also emphasises strengthening the open-source AI ecosystem through high-quality datasets, domestic model deployment, and collaborative innovation as a more viable path to strategic capability.

This approach also extends to partnerships. India’s collaborations, such as those aligned with initiatives like Pax Silica to make semiconductor supply chains resilient, reflect a model based on negotiated interdependence.

At the same time, deeper structural gaps remain. Long-term competitiveness, Reddy argues, will depend on expanding R&D investments and strengthening talent pipelines across academia and industry, areas that continue to lag behind global benchmarks.

While the central government increased the allocation for core R&D expenditure for key science departments in Budget 2026, it still accounts for 0.7% of the GDP—significantly lower than top innovators like South Korea, the US, Germany, and China. 

The IP vs Manufacturing Debate

One of the most critical fault lines in India’s hardware ecosystem is the distinction between manufacturing and intellectual property.

Gupta draws a sharp line between the two. “I would rather optimise more for owning the IP, rather than just putting things together,” he says.

He illustrates this with a common scenario: companies assembling imported components and branding them as “Made in India.” While this may create manufacturing activity, it does little to build true technological independence.

“Technically, you’ve done manufacturing in India… but actually the IP doesn’t lie with you,” he explains.

This distinction matters because IP ownership determines long-term value. Without it, Indian companies remain dependent on foreign suppliers for design changes, upgrades, and innovation. Yet building IP-led hardware companies is significantly harder than scaling manufacturing. It requires deep technical expertise, long gestation periods, and patient capital.

Beyond demand and IP, there is a third constraint: ecosystem depth.

Mohith Gauri, counsel at Axiom5 Law Chambers, points out that hardware companies need more than factories to succeed. They require a full-stack ecosystem like testing facilities, logistics networks, tooling infrastructure, and after-sales support.

“The biggest gap is ecosystem depth,” he says, adding that startups will only switch to domestic suppliers when hardware is “reliable, competitively priced, easy to procure, and supported” across the lifecycle.

Rethinking Make in India

Faced with weak domestic demand, many Indian hardware startups are looking outward.

Drone maker ideaForge recently secured an order to deploy its unmanned aerial systems for school security in the US. Meanwhile, quantum security pioneer QNu Labs partnered with US-based Viavi Solutions last year to develop network intelligence technologies.

Gupta suggests that entering global markets may be a more viable path, though challenging. One strategy is to start small, supplying components rather than full systems, and gradually build credibility. 

But it raises a fundamental question about the future of India’s “Make in India” ambitions in hardware.

If manufacturing depends on client scale, and clients themselves struggle to scale, then supply-side incentives alone may not be enough. Demand creation, through government procurement, anchor customers, and ecosystem support, becomes equally important.

Reddy’s assessment offers a more pragmatic path forward. Instead of chasing full-stack sovereignty, India should focus on building selective strengths, leveraging open ecosystems, and embedding itself strategically within global value chains.

India’s sovereign hardware ambitions are not misplaced, but they are incomplete. Hardware cannot scale in a vacuum. It is only as strong as the ecosystem around it.