INDIA: India is assembling an industrial-policy toolkit that includes production-linked incentives, the IndiaAI Mission, semiconductor subsidies, and lessons drawn from global innovation systems. The instinct is understandable. Governments want to compress technological catch-up through coordination and capital. Yet the lesson India appears to be drawing is more complicated than either its admirers or critics suggest.
The most consequential Chinese technology outcome of this decade was not produced by the Chinese state in the way it is often assumed. DeepSeek, the AI firm whose low-cost frontier models unsettled Silicon Valley and reshaped assumptions about the price of intelligence, did not emerge from a national champion program. It was not a state-picked winner under a five-year plan. It did not originate inside China’s formal industrial policy machinery.
Also Read: How Shivaani Bhardwaj Turned a Small-Village Dream into Digital Influence
That single fact complicates the standard account of how innovation is produced in China.
The conventional narrative describes China as the world’s most integrated system of state-directed innovation. Government-backed industrial funds, strategic blueprints such as Made in China 2025, and tightly coordinated consortia are often seen as the core drivers of technological progress. Scholars like Barry Naughton describe this as a re-engineering of the innovation chain, where the state shapes not only incentives but the structure of production itself.
Yet a closer look shows that much of this architecture exists to correct distortions created elsewhere in the system. China’s private sector has long been its most dynamic source of productivity growth, but state-dominated banking channels capital disproportionately toward state-owned enterprises. Industrial guidance funds and subsidies often function as corrective mechanisms, redirecting investment toward politically prioritised technologies rather than purely market-driven allocation.
In sectors with clear metrics and fast feedback loops such as electric vehicles, batteries, solar power, and telecommunications, this model has delivered scale and speed. But the record is uneven. The semiconductor “Big Fund” has faced corruption investigations. Failed ventures such as Wuhan Hongxin Semiconductor Manufacturing Company illustrate how large-scale state capital can be misallocated. Even the electric vehicle sector, despite its global leadership, is now dealing with intense price wars and overcapacity.
The deeper constraint is structural. China devotes a relatively small share of its research spending to basic science, the kind of open-ended inquiry that produces unexpected breakthroughs. While the United States, South Korea, Japan, and France allocate significantly higher shares to basic research, China’s system has historically favoured applied, near-term engineering outcomes tied to policy targets. This bias is now widely acknowledged within Chinese policy circles, even as efforts are underway to rebalance it.
This is where the apparent paradox of DeepSeek becomes important. If state direction tends to suppress or crowd out foundational research, how did a breakthrough firm emerge at the frontier of artificial intelligence?
The answer is that it did not emerge because the system selected it. It emerged because the system built the substrate beneath it. Dense engineering talent pipelines, large-scale compute infrastructure, and industrial clusters capable of rapid scaling formed the environment in which outlier firms can appear. The state did not pick the winner. It built the conditions under which a winner could unexpectedly emerge.
This distinction is central for India’s own policy direction.
India’s current strategy, anchored by the IndiaAI Mission, is often mischaracterised as replication of China’s champion-picking model. In practice, its structure is closer to an open-platform approach. Its compute allocation system distributes subsidised access to GPUs through competitive applications, with hundreds of proposals evaluated rather than a handful of firms selected. Similarly, semiconductor support under the Design Linked Incentive framework has been spread across a cohort of startups rather than concentrated in a single national champion.
Early indicators from India’s semiconductor design ecosystem suggest modest but tangible progress, including multiple tape-outs, initial chip fabrication runs, and early venture capital inflows into hardware startups. The direction of travel is toward ecosystem building rather than industrial centralisation.
However, a harder issue sits beneath the design logic: execution capacity.
Parliamentary disclosures show that actual budget releases for India’s AI and semiconductor programs have lagged significantly behind revised estimates in consecutive fiscal years. The gap is not merely administrative timing. It reflects a persistent pattern in Indian industrial policy: ambitious announcements, careful design, and incomplete funding.
This matters because open systems only work when they reach scale. India’s digital public infrastructure succeeded because it combined elegant architecture with sustained investment until network effects became self-reinforcing. Artificial intelligence infrastructure is vastly more capital intensive, and without consistent funding, open-access systems risk remaining underpowered.
A second structural issue lies in credit allocation. India’s banking system continues to favour collateral-backed lending, which disadvantages intangible-asset-heavy startups. While venture capital participation is increasing in chip design and AI, early-stage risk capital still depends heavily on policy signals rather than institutional depth.
A third challenge concerns basic research. India’s research and development intensity remains low by global standards, and its share devoted to foundational science is even smaller. This is the most important long-term constraint. Systems optimised for deliverables tend to underinvest in the unpredictable research that produces step changes in capability. Breakthroughs like frontier AI models rarely come from planned outputs. They emerge from environments that tolerate long uncertainty horizons.
The global lesson is not that China’s model should be copied or rejected. It is that it should be correctly understood. The Chinese state is effective at building industrial substrate at scale, but it does not reliably “select” innovation outcomes. The most important innovations often emerge from the edges of that system rather than its centre.
For India, the strategic question is therefore not whether to adopt industrial policy, but what kind of industrial policy architecture to sustain. Open systems can generate large-scale innovation, but only if they are adequately funded, institutionally protected, and allowed to support long-horizon research rather than only near-term production targets.
The risk is not conceptual confusion. It is execution drift: a policy framework that is intellectually aligned with modern innovation systems, but materially underpowered to make them work.
In the emerging global competition over compute, models, and semiconductor capacity, the countries that succeed will not simply be those that pick the right firms. They will be those that build the widest and deepest fields in which unexpected firms can win.
Also Read: Sagar Gorde Journey from Rural Struggles to a Street Food Vendor
