This is the result of what the industry calls “secondary” innovation. This is a technology that cannot be patented globally because it will affect global economics in the long term. The spending on basic engineering, research and development (ER&D) work in AI is huge, with at least five executives involved in AI-related work said mint.
In November, the World Intellectual Property Organization (WIPO) annual report stated that India is the sixth region in the world in terms of overall patent applications for China, the US, Japan, Korea and the European Union. However, the gap was severe. China filed 1.7 million patents in 2024, nearly three times more patents than the US, with 600,000 patents. India has filed only 90,000 patents. This is 5% of what China did.
Read again: AI companies with GPU SOPS may see government at the table
The gap is even more evident in Generation AI, the central battlefield of global technology now. Last year, China filed over 38,000 patents with Generic AI with WIPO, the global patent office, prior to the US, which has around 6,500 patents. India was also ranked 6th here, with 1,350 patents granted patents with the generation AI. 3.5% of China's progress, about a fifth of the US.
Last month, Minister of Electronics, Ashwini Vaishno promised, “India's first basic AI model is still expected to be released by the end of this year.” However, the patent application suggests that the US-China war over AI hegemony threatens to leave India from the League of Nations, which will affect global innovation and the economy over the next decades.
Lack of funds
The founders argue that much of this is because there is no large initial funding. Founded by Ashish Vaswani, a former Google Brain engineer who co-invented the trans model that favors all generative AI applications, US Essential AI emerged from stealth in December 2023 with a $56.5 million Series A funding round.
Others who have raised large U.S. capital over the past three years include the $65 million funding round for Adept AI in April 2022 and the $60 million Series A in August. Each of these ventures is currently investing in the construction of patented and licensed basic technologies to run AI applications and services worldwide in the long run.
Read again: Sovereign Silicon: India is targeting indigenous 2NM, Nvidia-level GPUs by 2030
Global venture leaders agree that India is behind the AI curve at this point.
“There is definitely a shortage of sufficient AI engineers working on core engineering in the Indian field,” said Pranav Mistry, founder and CEO. Mistry, former global chief of Samsung's advanced research department,mintBystanders at a gathering in Bengaluru earlier this month.
“There is certainly a difference in thinking between India and the US in terms of how ventures approach AI engineering in both countries. Ultimately, being able to hold patents provides access to geographical soft power over the next few years, and India will definitely need to focus on this sector.”
Vaswani of Essential AI said, “There is no reason why India doesn't build its own AI model. There should be more ventures focused on doing that in India and within India.”
Vision development
Investors argue that the long-term lack of vision from the founders is a key part of why Core ER&D jobs are not found among Indian AI startups.
“Entity pitching to implement basic AI engineering comes with a five-year roadmap, which comes with a five-year roadmap equivalent to decades in the modern AI world. It's absolutely true that India is building on top of the engineering that the US and other entities take on. Venture Capital Firm, Delflexor Ventures.
But the lack of funding is also an important reality. In India, there were no major early stage investments in AI-focused startups, except for Salvam's $41 million Series A funding round in December 2023. Noida-based Gan.ai and Bengaluru's Gnani.ai are two startups first supported along the Centre's $1.2 billion Indian mission, raising $5.25 million and $4 million respectively.
Gurugram-based Soket AI Labs, fourth of the first government-backed startup, has yet to raise a venture capital round and so far has only “approximately $3 million from Angel Investors.”
Government support
“This is why the government's AI mission to reduce the cost of access to processors for training AI models is so important, and we are pleased to provide fairness to the government in exchange for access,” Upperwal said.
Read again: The brain behind the AI generates his vision is set in India
“Indian venture capital investors have a limited appetite to invest in deep technology R&D. This is important for building new underlying AI architectures that AI startups can patent for global use in the long term and licensed for global use.
Policy experts say the issue goes beyond startups alone.
Startups are “as possible as the entire ecosystem and cannot solve fundamental problems across the industry,” said Rohit Kumar, founding partner of Quantum Hub and consultant for various government and public sector initiatives.
“Essentially, India's R&D is not yet fully prioritized. The budget is too small. The agency has no way of pursuing basic innovations for its US and Chinese counterparts,” Kumar said. “Incubators at top engineering institutions are hampered by bureaucratic processes not seen internationally. India is heavily bound in these ways.”
But in the long run, investors believe that the key balance between co-innovation and clever application development will be the right way. Vishesh Rajaram, managing partner of Speciale Invest, a venture capital firm focused on Deep Tech, said India “is a little behind in the curve at this point, but we haven't yet missed the AI bus.”
“Many basic jobs are difficult and there are multiple challenges to the story. Access to infrastructure is limited, and there is limited talent to actually get the underlying work and patents. As a result, there is room for startups to catch up, even from the core engineering efforts.
Prayank Swaroop, partner at Venture Capital Firm Accel, said for startups, “The real opportunities lie in dedicated AI applications that solve specific problems of scale. We can see Indian startups creating target solutions using existing foundation models as building blocks.
But others believe that more weight is the need for India's time due to basic innovation. Kumar from Quantum Hub cited advances in China as an example.
“We also need to capture a large amount of low margin secondary innovation market, but as China has proven, the profits that can be obtained from large-scale innovations need to be reinvested in basic innovation,” he said. “China is a clear example of how it works and this needs to be replicated more efficiently in India.”
