The US government’s recent directive restricting Anthropic from allowing its foreign customers to use Claude’s most advanced models reset a pecking order. Where India stood in that order could be gleaned even earlier from the now-rescinded Artificial Intelligence Diffusion Rule that the US Bureau of Industry and Security had issued in January 2025. It placed India on the second tier of access, clubbing it with third-tier China for validation of end use. Hence, it has been clear for a while that India needs to seriously re-evaluate its strategies for accessing state-of-the-art technologies and for developing tech sovereignty in a restricted landscape.
Even when provided whole-heartedly, tech transfers have traditionally not gone past making a version of the product the transfer came with. For example, India received knowhow to make Mig-21 fighter jets, but still needed to buy Mig-25, Mig-27 and Mig-29. Tech transfers for Mig-29 and Sukhoi-30MKI played a minimal role in getting the indigenous Kaveri jet engine off the ground. This shows that in the name of transfer of technology, India has ended up getting only transfer of manufacturing.
Tech transfers are never complete, as tacit knowledge is almost impossible to transfer. A healthy transfer would allow subsequent advanced versions be built from the learnt knowledge. Thus, the best way to procure a foreign technology is to buy the company or business unit that owns the technology.
Examples for this do exist. During the pandemic days, Rolls-Royce sold its stake in the Spanish ITP Aero, which makes jet engines for Rafale’s competitor, the Eurofighter Typhoon. Along with other capabilities, ITP Aero also built the Eurojet EJ200 engine, an upgraded variant of which (EJ230) it bid for the Tejas Mark II project, though it lost the order to General Electric’s F414. Eventually, ITP Aero was bought for €1.7 billion—a pittance in the context of its strategic importance and the industry—by an American private equity group.
One might argue: why buy a sinking business when one just needs the technology? History shows that by moving labour-intensive manufacturing to India from higher-cost territories, Indians have been able to turn around many businesses. Tata’s acquisition of Jaguar Land Rover comes to mind.
Another tech access strategy frequently talked about in bureaucratic circles is hiring foreign high-tech workers like those available during the fall of the Soviet Union. The private sector has a recent success story in telecom, where a large player hired an army of expats to build their infrastructure.
Another tech procurement strategy—seen more as a taboo in India—is reverse engineering. Due to its low acceptance, those who dabble in it are susceptible to the Dunning-Kruger effect, or the lack of foundational knowledge that’s required to realise what one doesn’t know. Reverse engineering continues to be misunderstood as a theft of intellectual property, though a landmark US Supreme Court judgement provided legitimacy for its use. In Bonito Boats vs Thunder Craft (1989), reverse engineering was adjudged “the very lifeblood of a competitive economy”. In Kewanee Oil vs Bicron (1974), the US Supreme Court defined reverse engineering as “starting with the known product and working backward to divine the process which aided in its development or manufacture”.
Limits exist through patents and contractual laws, but they do not cover the whole ground. For example, Indian IP laws provide safeguards that are used by the pharmaceutical industry. It’s time the Indian jet engine industry learnt from it.
Almost all primary jet engine manufacturers have, at some stage, dabbled in reverse engineering to discern a technology legitimately. When done within a company, reverse engineering is also called re-engineering. Most jet engines today are built through such methods.
General Electric very proudly displays its “Hush-hush Boys”, a top-secret engineering team assembled during World War II, on their website. What was the secrecy all about? Hindustan Aeronautics Limited (HAL) provides the most fertile ground for such an endeavour. Over decades, the company has worked with a large variety of jet engines from all over—from completely knock-down kits, and semi knock-down ones to those needing far more assembling. Though HAL does have R&D centres, it has not ventured to wholeheartedly compete with DRDO’s Gas Turbine Research Establishment.
Reverse engineering needs to be incorporated in the Indian technical education system. A separate collaborative stream can be taught at all levels starting with BTech. The use of such skills can reduce development times for complex systems by up to 80 percent and for simpler ones by about 60 percent.
Unfortunately, many believe in the toolbox fallacy in technology development—the self-deception that one cannot achieve a goal, start a project, or learn a skill until they acquire the ‘perfect’ tools for it. This leads to grandiose budget allocations for exotic tools and infra, which often settle into centres of excellence. More often than not, these investments lead to a false sense of progress.
As we stay shy, China and Iran perfect the artful science to meet their needs. Iran’s hold-out against the world’s most powerful army with cheaply built Shahed-136 drones provide the counter that technology can indeed be built in the most challenging of environments. That the US itself reverse-engineered the Shahed-136 into its LUCAS (Low-cost Uncrewed Combat Attack System) drone confirms the premise! India holds many such opportunities for frugal technology. For example, Mohali’s Semi-Conductor Laboratory has the technology for making global navigation satellite system receiver chips, including for India’s NavIC.
Circling back to AI, India is still way back in the queue for access. Large language models and the advanced chips needed for making them seem elusive in the near future. Given the limited hardware access, a better route might be through small language models, which can be trained to indigenous needs and ply on cheaper infrastructure.
It is quite possible to achieve the dream of a developed nation in the next two decades. But India cannot afford to take the scenic route for that. It will need to buy companies, learn to reverse-engineer, shun the toolbox fallacy, be more welcoming of expat talent, invest in frugal innovations and take the SLM route to a variety of specialised models—all on its way to tech sovereignty.
Anshuman Tripathi | Tech evangelist, Sloan Fellow at Stanford Graduate School of Business and former member of National Security Advisory Board
(Views are personal)