India is no longer content with being the world’s IT back office
At the India AI Impact Summit, New Delhi is making an unambiguous pitch: India doesn’t just want to use AI, it wants to host the data centers, train the models, and set the rules of the game. From $100 billion infrastructure plans to new OpenAI and Anthropic offices, the country is moving at a speed that should worry both Silicon Valley and Brussels.
In this piece, we’ll unpack what was actually announced, why Big Tech and chip makers are piling in, what this means for the traditional Indian IT model, and why Europe – which sent Emmanuel Macron but not much capital – cannot afford to ignore India’s new AI ambitions.
The news in brief
According to TechCrunch’s reporting from the India AI Impact Summit in New Delhi, India is using a four‑day mega‑event (around 250,000 visitors) to roll out a full AI industrial strategy.
Key announcements include a $1.1 billion state‑backed VC fund for AI and advanced manufacturing startups, and Adani’s plan to invest $100 billion in renewable‑powered AI data centers by 2035, expecting to pull in a further $150 billion of adjacent investments.
OpenAI said India now has over 100 million weekly active ChatGPT users, second only to the U.S., and will open offices in Bengaluru and Mumbai, plus deploy up to 1 gigawatt of compute with Tata. Anthropic is opening its first Indian office in Bengaluru and partnering with Infosys on enterprise deployments.
Multiple Indian startups launched or updated large language models and speech technologies in Indian languages, while the government’s BharatGen consortium unveiled a 17B‑parameter model for 22 local languages. India also joined the U.S.-led Pax Silica supply‑chain group and announced that over 88 countries and organizations signed the New Delhi AI declaration on using AI for social and economic good.
Why this matters
India is effectively declaring that it wants to be one of the world’s main compute and talent hubs for AI – not just a low‑cost outsourcing destination. The winners here are obvious:
- Big AI labs and cloud providers gain direct access to a vast, young, English‑speaking user base and a government keen to subsidize infrastructure.
- Local infrastructure and conglomerates such as Adani, Tata, and major IT firms get a once‑in‑a‑generation capex wave in data centers, energy, fiber, and chip‑adjacent industries.
- Indian AI startups suddenly have domestic access to serious compute and capital instead of being forced to domicile in the U.S. to scale.
The potential losers are just as clear. India’s traditional strengths – IT services and BPO – are precisely the sectors that Vinod Khosla warns could “almost completely disappear” in five years due to AI automation. When HCL’s CEO says Indian IT will focus on profits rather than job creation, that’s a quiet admission: the old “millions of entry‑level coders” model is running out of runway.
For the rest of the world, this summit is a signal that AI’s geography is shifting. Compute is clustering not only in the U.S. and China, but also in the Middle East and now India. Countries that don’t move quickly on infrastructure, skills, and regulation risk becoming just AI importers – paying rent to run foreign models in foreign clouds.
The bigger picture
What India did this week fits into three converging trends.
1. The global scramble for exascale compute.
The G42–Cerebras deal to deploy 8 exaflops of compute in India, plus OpenAI–Tata’s 1‑gigawatt plan and Adani’s $100 billion bet, show that AI infrastructure is becoming a new category of strategic asset, much like LNG terminals or 5G networks. Countries are no longer comfortable relying solely on U.S. hyperscalers; they want compute that is geographically and politically closer.
2. From AI Safety Summits to AI Industrial Summits.
Where the UK’s Bletchley Park summit in 2023 focused on existential risk and model governance, India’s event is unapologetically about growth: capital, jobs, and exports. The New Delhi declaration emphasises AI for development, while India simultaneously joins Pax Silica to secure materials and chips. It’s a reminder that, outside the transatlantic bubble, the main AI question is not “Is this too powerful?” but “How fast can we get a slice?”
3. The rise of non‑Western model ecosystems.
Government‑backed models like BharatGen’s Param 2, Sarvam’s open‑source 30B and 105B models, and multilingual offerings from Cohere Labs show that the foundation‑model world will not be monopolised by U.S. giants. They also validate a playbook Europe has talked about but rarely executed: build strong local models for under‑served languages, then push them into devices and verticals (phones, cars, education).
Taken together, the summit underscores that AI is now classic geoeconomics: compute, energy, language, and regulation intertwined.
The European / regional angle
Europe is present in New Delhi mostly via people – Macron on stage with Modi – rather than through big cheques or infrastructure commitments. That contrast is striking.
While India announces tens of billions in data‑center capex, the EU is still wrestling with how the AI Act will apply to foundation models and what “systemic risk” thresholds should be. Europe is building an advanced regulatory framework but not yet an equally ambitious compute strategy.
For European companies, India’s AI push cuts both ways:
- Opportunity: Cheaper, large‑scale compute in India (especially if genuinely renewables‑powered) could become an attractive option for European startups and enterprises, provided GDPR‑compliant data residency and contractual safeguards are in place. Partnerships with Indian players on multilingual models could also benefit Europe’s own under‑resourced languages.
- Risk: If European regulators move faster than investors, Europe may end up consuming AI built and hosted elsewhere. The fact that India joined the U.S.-led Pax Silica supply chain group and not an EU‑driven initiative shows who is currently setting the industrial agenda.
European privacy culture and the DSA/DMA framework will also make EU firms more cautious about routing user data through non‑EU infrastructure. That opens a gap: either European cloud providers tie up with Indian partners under strict data‑protection contracts, or they watch U.S. and Asian players dominate an India‑centric AI stack.
Looking ahead
Expect three things over the next 24–36 months.
1. A brutal skills reshuffle inside India – and more competition globally.
As AI tools become embedded in IT workflows via partnerships like Anthropic–Infosys or OpenAI–Tata, mid‑level IT and BPO roles will come under pressure. India will try to move millions of developers “up the stack” into AI product building, not just services. That means European and U.S. startups will face even fiercer competition from Indian teams that are faster and cheaper, but now also working with frontier‑grade tools.
2. A race to prove the sustainability story.
Sam Altman’s dismissive framing of concerns about AI’s water and energy usage will not land well in Europe, where regulators and NGOs are already scrutinising data‑center footprints. If India wants to be the world’s AI server room, it will be forced to demonstrate credible renewables, cooling, and grid‑stability strategies. Watch for joint announcements involving European utilities, turbine makers, or green‑finance institutions.
3. More AI geopolitics, fewer neutral players.
By joining Pax Silica and hosting exascale compute from Middle Eastern and U.S. partners, India is quietly locking itself into certain supply chains and alliances. Europe will have to decide: does it deepen AI ties with India as a counterweight to both the U.S. and China, or does it treat India mainly as a market and outsourcing location?
Unanswered questions remain. Will India develop a coherent domestic AI regulatory framework beyond broad declarations of “AI for good”? Will its promised $200 billion in AI‑infra investments actually materialise? And crucially, will European policymakers recognise that compute strategy is now as important as AI rules?
The bottom line
India’s AI Impact Summit is less about press releases and more about positioning: the country is bidding to become a primary global hub for AI compute, talent, and language technologies. If even a fraction of the announced investments land, the old map of “AI equals U.S. plus China” will look outdated.
For Europe, the choice is stark: engage with India as a strategic AI partner – on infrastructure, models, and regulation – or watch another continent build the platforms on which European innovators will have to rent space. The question for readers is simple: in this new AI geography, do you want to own part of the stack, or just an app icon on top of it?
