Zoho co-founder Sridhar Vembu has reignited a strategic debate in India's technology ecosystem by arguing that the country should prioritise tech sovereignty, infrastructure and foundational capabilities over large-scale spending on GPUs. In a post that reflects a long-standing concern about dependence on foreign technology stacks, Vembu said India's path to lasting prosperity lies in building core capabilities rather than rushing to match global AI leaders through expensive hardware purchases.
The remarks come at a time when governments, startups and investors across the world are racing to secure access to advanced computing power, especially graphics processing units that are essential for training and running modern artificial intelligence systems. For India, the temptation to invest heavily in GPUs is understandable: the country wants to remain competitive in AI, support domestic startups and reduce the risk of being left behind in a fast-moving global market. But Vembu's intervention suggests that such spending may not deliver the strategic autonomy India needs if it is not paired with deeper investments in infrastructure, engineering talent and indigenous technology development.
Sovereignty Over Hardware
Vembu's central argument is that technological strength cannot be reduced to the possession of expensive chips alone. GPUs are important, but they are only one layer in a much larger stack that includes power supply, data centres, networking, cloud infrastructure, software frameworks, semiconductor supply chains and the ability to design and deploy systems independently. Without those foundations, he implied, India risks becoming a consumer of foreign-built capabilities rather than a creator of its own.
That view aligns with a broader school of thought in India's technology policy debate, which holds that the country should avoid over-indexing on headline-grabbing AI infrastructure purchases and instead invest in the less glamorous but more durable building blocks of industrial strength. These include reliable electricity, domestic manufacturing capacity, research institutions, open-source software ecosystems and the training of engineers who can build and maintain complex systems at scale.
Vembu has often spoken about decentralisation, self-reliance and the need for India to develop its own technological muscle. His latest comments extend that philosophy into the AI era, where the global competition is increasingly shaped by access to compute. The concern, from his perspective, is not merely financial. It is strategic. If India depends too heavily on imported GPUs and the foreign firms that control them, then the country's AI ambitions may remain vulnerable to supply constraints, export controls and pricing power set abroad.
The AI Spending Dilemma
The debate is especially relevant for India's startup sector, where founders are under pressure to incorporate AI into products, services and workflows. Many companies are eager to access large-scale compute, but the economics are difficult. GPUs are costly, scarce and often concentrated in the hands of a few global cloud providers. For a country with deep ambitions but finite capital, the question is whether it is wiser to spend heavily on imported compute or to channel resources into building the conditions that make such compute more productive over time.
Vembu's position suggests that India should resist the instinct to equate technological progress with capital expenditure on the most visible components of the AI stack. Instead, he is arguing for a longer horizon: one in which India develops the infrastructure and institutional capacity to innovate on its own terms. That would mean not just buying hardware, but creating the ecosystem that can design chips, host models, train talent and support a broad base of domestic innovation.
The message also carries political and economic weight. India has repeatedly framed self-reliance as a national priority across sectors ranging from manufacturing to digital public infrastructure. In technology, however, the challenge is more complex because the most advanced layers of the stack are deeply globalised. Vembu's comments effectively ask whether India should accept that dependence as a temporary reality or confront it as a strategic weakness that must be reduced over time.
Building Lasting Capacity
For policymakers, the implication is that AI strategy should not be measured only by the number of GPUs acquired or the size of a national compute cluster. It should also be judged by whether those investments strengthen domestic capabilities in the long run. That includes whether Indian firms can build competitive products, whether researchers can access the tools they need, and whether the country can reduce exposure to external chokepoints in critical technologies.
For startups and venture investors, the remarks are a reminder that the next phase of India's technology growth may depend less on chasing the most expensive infrastructure and more on building resilient systems that can scale efficiently. In that sense, Vembu's intervention is not an argument against AI, but against a narrow definition of progress. His message is that prosperity will come from owning the foundations, not merely renting the latest hardware.
As India positions itself in the global AI race, the tension between speed and sovereignty is likely to intensify. Vembu's comments place him firmly on the side of sovereignty: a belief that the country should invest first in the infrastructure, institutions and technologies that can outlast any single hardware cycle. In a sector often driven by hype and capital intensity, that is a notably disciplined and strategic warning.