FOR some time now, India’s tech industry has been reluctant to plainly say an uncomfortable truth: the engine that powered three decades of economic growth is sputtering. Tata Consultancy Services, India’s largest IT employer, shed over 23,400 jobs in the year ending March 2026. Infosys cut the headcount by nearly 20,000. Close to 40,000 layoffs have hit the sector in a year. The culprit is not a recession. It is artificial intelligence (AI), the very technology India’s leaders now race to harness.
Prime Minister Narendra Modi’s recent tour to the UAE, Netherlands, Sweden, Norway and Italy was shaped around this reckoning. It is pertinent to understand why each agreement matters and what the country is up against as the tech ecosystem changes.
India’s IT services sector was built on a simple proposition: hire large numbers of engineers, train them cheaply and sell their labour to western corporations. It worked brilliantly for a generation until generative AI began automating the routine coding, testing and back-office work that formed that model’s bedrock. Agentic AI is projected to redefine over 10 million Indian jobs by 2030. The traditional pathway — graduate, join a mid-tier firm, bill hours to a US client — is narrowing fast.
India can’t wait for AI to create replacement jobs. It must own parts of the global AI economy, not just service it. That means building the hardware that AI runs on, infrastructure that powers it and research pipelines that advance it. Each stop on Modi’s tour addressed one of these gaps.
The most impactful announcement came in Abu Dhabi. Modi and UAE President Sheikh Mohamed bin Zayed Al Nahyan oversaw the signing of an agreement between the G42 Group, the Mohamed bin Zayed University of Artificial Intelligence, and India’s Centre for Development of Advanced Computing (C-DAC) to establish an 8-exaflop supercomputing cluster in India which will be among the most powerful computing infrastructure anywhere on the planet.
AI is fundamentally a compute problem. The more powerful your infrastructure is, the more sophisticated the models you can build. India’s researchers and startups have historically rented expensive computing from US cloud providers like Microsoft, Amazon and Google. This dependence limits what they can build and who controls it. The UAE deal gives India sovereign computing muscle to train AI models domestically, in Indian languages, for Indian priorities, rather than adapting tools built for western markets.
Every AI model, smartphone and electric vehicle runs on chips. India has world-class chip designers at Qualcomm, Intel and NVIDIA, but India has never manufactured its own chips, remaining dependent on Taiwan and South Korea. This vulnerability was exposed sharply during the Covid-era chip shortage.
In The Hague on May 16, Tata Electronics and ASML, the Dutch company holding a global monopoly on lithography machines without which no advanced chip exists, signed a landmark MoU. ASML will supply equipment and expertise to Tata’s $11 billion, 300-mm semiconductor fabrication plant in Dholera, Gujarat. This is India’s first commercial front-end chip fab, covering workforce training and supply chain development too. Semiconductor fabrication is one of the most complex manufacturing endeavours in human history. The Dholera plant will produce chips specifically for AI, automotive and mobile applications.
The old IT model relied on quantity, with large numbers performing standardised work. The AI economy needs quality: researchers and innovators who can build, not just execute. In Stockholm on May 18, India and Sweden elevated ties to a Strategic Partnership and launched the India-Sweden Technology and Artificial Intelligence Corridor, alongside Joint Innovation Partnership 2.0, targeting doubled bilateral trade within five years.
The AI corridor is the critical piece. Sweden, home to Ericsson and world-class 5G/6G research, offers the deep-tech collaboration India needs to upgrade its talent profile. Indian engineers gain access to Swedish research institutions. Swedish firms gain India’s market and scale. India’s 6G ambitions gain a credible industrial partner in Ericsson.
AI has a power problem that receives insufficient attention. Data centres running AI models are among the world’s fastest-growing electricity consumers. As India scales its ambitions, including the 8-exaflop cluster, it needs massive clean power. At the 3rd India-Nordic Summit in Oslo, leaders elevated the India-Nordic relationship to a Green Technology and Innovation Strategic Partnership. Norway’s hydropower and Denmark’s wind make Nordic nations natural partners for powering India’s data infrastructure cleanly. Norway’s interest in India Stack — the UPI and Aadhaar-based digital public infrastructure — also signals something significant: India’s homegrown digital model is becoming an exportable asset, a tech product India can offer to the world.
Italy may seem an unlikely technology partner, but its strengths in aerospace, advanced engineering and industrial robotics are directly relevant. In Rome, on May 20, India and Italy elevated ties to a Special Strategic Partnership, with AI and clean energy at its core. As India seeks to automate factories and climb the value chain in electronics and automotive manufacturing, Italian industrial expertise is a practical bridge, a relationship further deepened by the landmark India-EU Free Trade Agreement.
If one sets these agreements side by side, a coherent strategy emerges. India is betting that AI’s disruption of its IT model is not just a threat but a mandate to rebuild at a higher level — manufacturing the chips AI needs, owning the computing infrastructure it runs on, building research corridors that produce AI talent, powering infrastructure cleanly and embedding itself in the manufacturing networks that AI is transforming.
This is the right bet. But agreements are not outcomes. Jobs are still being shed. The Dholera fab will take years to yield its first chip. Political will must become industrial reality — and faster than AI is moving. That is India’s real test. So much work has to be done to implement more prudent measures to scale up the tech ecosystem value chain. These agreements mark the first steps of many consistent steps that need to be undertaken, and fast.