What’s Stopping India from Its Own DeepSeek Moment?

India’s IT and tech ecosystems are often applauded for their resilience, innovation, and impact at a global level. Nonetheless, when it comes to deep-tech breakthroughs and foundational AI research, the country continues to lag behind global leaders like the US and China. The recent emergence and success of DeepSeek, a low-cost, open-source AI model pioneered in China, crudely reminds India of the structural challenges it should overcome to achieve its own “DeepSeek moment.” 

Let’s be brutally honest: India has surely come a long way in IT services and outsourcing but is far away from matching China's prowess in groundbreaking innovations and cutting-edge technologies. More investments in research and a greater number of AI researchers are needed, at the very least, to bring existing talent out of obscurity.

For this, we need to overcome issues like lack of R&D investments, building world-class state of innovation hubs, and addressing employment woes. I agree that we have the largest student base compared to any other country, but specialization is what counts more than mere headcounts.

The Funding Gap

One of the most glaring issues is India’s chronic underinvestment in research and development (R&D). The country spends a mere 0.7% of its GDP on R&D, compared to China’s 2.65%. Worse still is the private sector contribution, currently a miserable 0.28% of GDP, compared to 2.06% in China. This funding shortfall has a direct effect on India's capacity to innovate and compete on a global scale.

While China’s “Made in China 2025” initiative and its AI development plan are backed by massive government funding and clear targets, India’s AI mission is commendable but underfunded, with a budget of $1.25 billion (a drop in the ocean as compared to $150 billion, which is China’s commitment by the year 2030). 

Skill Gaps and Brain Drain

 There’s a huge skill gap in emerging technologies like AI, machine learning, and deep learning in the country despite India producing a huge number of engineers. The situation is further magnified by the outflow of India’s top AI talent — 60% of whom see better research opportunities and infrastructure in the U.S. Keeping this talent within the nation requires high salaries, top-notch labs, and an innovative culture — things that India still lacks. Without a robust talent pipeline, India’s AI ambitions will remain unfulfilled.

The Ecosystem Challenge

One reason for China’s success in AI is strong collaboration between schools and industry. In India, they are fragmented and underfunded. Our universities seldom make it to the global top 100 lists, and there is little emphasis on interdisciplinary research and experimental learning—both of which are vital for breakthroughs. Moreover, India’s tech ecosystem tends to be risk-averse — preferring incremental innovation over moonshot projects. DeepSeek’s success hinged on bold experimentation with cheap hardware and open-source models, areas in which India has yet to make big strides.

Data and Infrastructure: The Twin Bottlenecks

Digital-first datasets are among the most significant hurdles toward building a foundational (or general-purpose) AI model for India. Access to quality, labelled, and annotated data still continues to be a significant hurdle, particularly when we compare it with the developed countries that have more established data integrity. India needs to build its own datasets through community sourcing and use the data held by the government behind the firewalls for innovation. 

Equally critical is the issue of compute capacity. India's current data centre infrastructure is not geared to support the high computational loads of large-scale AI training. AI workloads require high-density racks (50kW+ per rack), and most of the current facilities are not configured to support these loads, which would require liquid cooling or immersion cooling solutions to deal with the heat effectively. The deployment of AI-capable data centres fuelled by clean, hybrid energy solutions is critical to support the growing demands of AI Research & Development.

Geopolitical Constraints

Adding to these challenges are the U.S. government’s export restrictions on GPUs sold to India, capped at 50,000 GPUs overall. This regulation affects Indian players, whereas U.S.-based hyperscalers can still buy an unlimited number of GPUs. This inequality will end up constraining the ability of Indian players. In order to counter this, India will need to spend on creating its own chips and GPUs, bringing down the reliance on foreign vendors and screening against future sanctions.

Way Forward

In spite of these setbacks, India can still attain its own DeepSeek moment. The initial step is to raise R&D expenditure to a minimum of 2% of GDP, with major private sector involvement. Encouraging industry-academia collaboration is equally vital. For instance, establishing 10+ research centers with state-of-the-art infrastructure by 2026 would be revolutionary. Along with that, the best talent needs to be lured with competitive offerings and world-class research infrastructure, whereas encouraging risk-taking and open-source innovation can help India leapfrog into the global AI race.

Furthermore, India also needs to prioritize building high-quality, India-centric datasets through community sourcing and investing in AI-ready infrastructure. The role of the government in facilitating access to data and in supporting domestic chip design cannot be overemphasized. Alongside, developing a framework for ethical AI practices and figuring out how Indian models are monetized will become crucial to ultimate success.

Even though the road ahead is tough, India's strengths—a huge pool of talent, the burgeoning startup sector, and a developing digital economy—are the bedrock to build AI innovation. The DeepSeek moment can be more than a wake-up call; it can be an inflection point for India to reboot its game in AI. By tackling the structural challenges aggressively, promoting collaborations, and fostering fearless experimentation, India can assert itself as the world leader in AI.

Now is the time to act. India's future with AI is not only about catching up—it's about pioneering.

By Sandeep Agarwal, India MD & Global CTO, Visionet Systems