Will India Build AI, or Just Feed It? The Five-Layer Challenge
At the AI Summit in New Delhi, Indian startups—Sarvam, Gnani, and BharatGen—showcased home-grown AI models. CEOs from leading technology firms praised India’s growing role in artificial intelligence and announced significant investment pledges.
Beyond the applause and investment pledges, a harder question remains: where does India stand in the global AI order? One way to assess its position is through the five-layer AI stack described by Nvidia chief executive Jensen Huang. Understanding India’s place in each layer reveals both the country’s strengths and its critical vulnerabilities.
The Five-Layer AI Stack
At the base is energy, which powers data centres and AI factories. AI requires reliable, high-quality electricity at competitive prices. India offers relatively low-cost power, but grid reliability issues, transmission bottlenecks, and uneven power quality remain constraints for hyper-scale infrastructure.
Next is compute infrastructure—advanced chips and high-speed networks. The US leads through Nvidia and major cloud providers, while China is rapidly expanding despite chip restrictions. India is growing its data-centre capacity but depends entirely on imported chips and GPUs, creating a strategic vulnerability.
The third layer is foundation models trained on vast datasets. The US leads through OpenAI and Google, while China has built strong domestic models such as DeepSeek. India’s strength lies in multilingual data and talent, but it lacks globally competitive models.
Above this are platforms and developer ecosystems. India benefits from a large developer base and digital public infrastructure, yet lacks globally dominant AI platforms.
At the top is the application layer, where India can build cost-efficient AI solutions across sectors. Its challenge is to capture intellectual property and build globally competitive products.
In short, despite its many strengths, India is not at the cutting edge of frontier AI.
Why Big Tech Is Interested
Why, then, are US technology firms focusing so intensely on India? Two reasons explain this attention.
First, India as the world’s largest open data source. China’s data is inaccessible to foreign firms. India produces vast streams of multilingual, real-world data from payments, e-commerce, mobility, agriculture, and public services. Some estimates suggest that Indian data volumes even exceed the US data used to train major AI models. For global AI companies, India is a vital testing and training ground for US models designed to serve the world.
US tech giants are investing in India to secure future continued supply of Indian data. The dynamic feels like this: US Big Tech fishes freely in India’s data pond, sells the cooked meal back at a premium—and if it causes harm, there is no accountability.
Second, India as the next billion-user AI market. ChatGPT alone has roughly 100 million weekly users in India, making it the platform’s second-largest market. With China largely closed to them, the next 500 million AI users for US firms are expected to come from emerging digital economies led by India. With hundreds of millions of smartphone users, a large developer base, and digital public infrastructure that enables rapid adoption, India offers unmatched distribution scale for US AI products.
At the AI Summit, Sriram Krishnan, Senior White House Policy Advisor on Artificial Intelligence, made the intentions clear. He said the US wants the world, including India, to use American AI models, infrastructure, and platforms.
Securing India’s AI Future
This requires action on three fronts.
First, developing Indian AI requires keeping Indian data available for domestic use. If data flows freely abroad, incentives to build local AI systems weaken. India’s automobile industry grew behind high import tariffs that encouraged global carmakers to invest and manufacture locally. A similar logic applies to data. If unrestricted data flows continue, Indian AI development may struggle to move beyond the small startup stage.
The February 6 India-US Joint Statement linked to the proposed Bilateral Trade Agreement calls for clear digital trade rules that could limit India’s ability to regulate its digital economy, bind it to free cross-border data flows, and prevent taxes on digital transactions. Agreeing to such provisions could reduce policy flexibility and public revenue as economic activity moves online.
Second, India must build its own AI models and infrastructure. Governments around the world are investing in national AI systems to secure technological sovereignty. China, France, South Korea, and the UAE are building models tailored to their languages, regulations, and local needs. India has begun this journey through the IndiaAI Mission and initiatives such as Sarvam, Gnani, and BharatGen.
Yet India’s AI startups, though driven by talented and enthusiastic founders, operate with limited capital and computing resources. Competing with global AI leaders requires sustained funding, access to high-performance compute, and long-term policy support—conditions that most startups cannot secure on their own.
Large Indian IT firms, with deep pockets and technical capability, are better positioned to build AI platforms. However, their business model—built on providing services to US technology companies—creates a conflict of interest. Developing indigenous AI tools could disrupt existing contracts and revenue streams, so many remain focused on processing data and supporting global platforms rather than creating their own models.
With domestic giants staying on the sidelines and global players seeking dominance, promising startups risk early acquisition, and India risks technological dependence. To avoid becoming merely a data supplier that pays to use tools built from its own data, the government must craft a strategy for large-scale investment, shared compute infrastructure, and long-term ecosystem support, as other countries are doing.
Third, prepare India’s IT workforce for the AI era. AI is automating routine coding, testing, and back-office tasks that support traditional outsourcing. While this will cut profits and jobs, it will also create major new opportunities. Integrating AI tools with legacy systems in factories, banks, hospitals, government platforms, and supply chains will keep Indian IT firms busy for years, as deployment requires customisation, regulatory compliance, cybersecurity safeguards, and human oversight.
Demand will also grow in cloud modernisation, data engineering, AI governance, and responsible deployment. This last-mile integration is where value will be created—and where India’s technical talent can lead globally. To capture these opportunities, India must strengthen engineering education, improve curricula quality, and prioritise skills for an AI-driven economy.
Conclusion: Rule-Maker or Largest Marketplace?
The New Delhi summit celebrated India’s AI potential, but the real test lies ahead. If India safeguards its data, builds its own models, and prepares its workforce, it can shape the AI age rather than merely consume it. Without decisive action, it risks supplying the data while buying back the tools built from it.
The choices made now will determine whether India becomes a rule-maker in the AI era—or simply its largest marketplace—with consequences for national security and economic sovereignty as well.
Q&A: Unpacking India’s AI Strategy
Q1: What is the five-layer AI stack and where does India stand in each layer?
The five layers are: energy (India has low-cost power but grid reliability issues); compute infrastructure (India depends entirely on imported chips/GPUs); foundation models (India lacks globally competitive models despite multilingual data); platforms and developer ecosystems (India lacks dominant AI platforms); and applications (India can build cost-efficient solutions but must capture IP). India is strong in the top application layer but vulnerable at lower, foundational layers.
Q2: Why are US tech firms so interested in India despite its lack of frontier AI capability?
Two reasons: India is the world’s largest open data source—producing vast multilingual data from payments, e-commerce, agriculture, etc., making it a vital testing/training ground for US models. Second, India is the next billion-user AI market—ChatGPT already has 100 million weekly users in India, and with China closed, India offers unmatched distribution scale for US AI products.
Q3: What is the “data sovereignty” concern raised in the article?
The article compares data to the automobile industry, which grew behind tariffs that encouraged local manufacturing. If Indian data flows freely abroad without restrictions, incentives to build local AI systems weaken. The India-US Joint Statement could limit India’s ability to regulate digital economy, bind it to free cross-border data flows, and prevent digital taxes—potentially reducing policy flexibility and public revenue.
Q4: Why aren’t large Indian IT firms building indigenous AI platforms?
Their business model is built on providing services to US technology companies. Developing indigenous AI tools could disrupt existing contracts and revenue streams. So they remain focused on processing data and supporting global platforms rather than creating their own models. This creates a conflict of interest that leaves promising startups to compete with limited resources.
Q5: What three actions does India need to take to secure its AI future?
First, keep Indian data available for domestic use through appropriate regulations. Second, build Indian AI models and infrastructure through large-scale investment, shared compute, and ecosystem support—as other countries are doing. Third, prepare the IT workforce for AI by strengthening engineering education and prioritising skills for AI-driven economy. Without decisive action, India risks becoming a data supplier that pays to use tools built from its own data.
