India AI Summit Bets Big on Billions, But the Real Challenge Lies Ahead

• India’s AI summit garners US$250+ billion in infrastructure pledges and new sovereign AI.
• Scaling AI beyond the summit floor presents a distinct challenge altogether.

When India hosted its AI summit at Bharat Mandapam in New Delhi last week, the optics were impossible to miss. French President Emmanuel Macron was in the audience. UN Secretary-General António Guterres addressed the crowd. Over 20 heads of government attended.

For five days, the country did not merely participate in the global conversation on artificial intelligence, but occupied the chair. And in doing so, it laid out a pitch that is as bold as it is complicated: that India, with all its linguistic diversity, informal economies, and scale, is the world’s best proving ground for AI that actually works.

The India AI summit, formally the AI Impact Summit 2026, is the fourth in a series of global AI gatherings following Bletchley Park, Seoul, and Paris – and the first to be hosted by a Global South nation. That distinction alone carried symbolic weight. But the numbers that came out of New Delhi over five days made clear this was more than symbolism.

A quarter-trillion dollars and counting

Infrastructure-related investment pledges from the India AI summit crossed US$250 billion, with an additional US$20 billion committed in venture capital and deep-tech investments, according to IT Minister Ashwini Vaishnaw.

India’s own conglomerates led the charge. Reliance Industries outlined a US$110 billion data centre and edge compute expansion programme over seven years. Adani Group announced plans to scale its data centre platform from roughly two gigawatts to five gigawatts at a cost of approximately US$100 billion.

Tata Group, meanwhile, said it would build AI-optimised data centres scaling toward one gigawatt, with OpenAI named as its first anchor tenant. Global tech giants reinforced the sentiment. Microsoft reiterated its US$17.5 billion commitment to cloud and AI infrastructure in India, part of a broader plan to invest US$50 billion in the Global South by 2030.

Google announced a US$15 billion investment in foundational AI infrastructure and a new America-India Connect initiative – fresh fibre-optic routes linking the US, India, and the southern hemisphere. The Indian government also confirmed plans to add 20,000 GPUs through its IndiaAI Compute initiative, with subsidised access for local startups at roughly a quarter of prevailing global market rates.

On the domestic AI development front, at least three new Indian models were launched during the summit. Sarvam AI introduced open-source models built on mixture-of-experts architectures, with multilingual and multimodal abilities designed for Indian languages. Government-backed BharatGen showcased Param 2, a 17-billion-parameter model supporting all 22 scheduled Indian languages.

Gnani.ai unveiled Vachana, a zero-shot text-to-speech system in regional languages. Global players also formalised their India commitments: OpenAI launched its “OpenAI for India” programme, and Anthropic announced an enterprise partnership with Infosys to support Claude deployments in Indian organisations.

Complexity as an asset

Beyond the headline figures, India’s core argument at the AI summit was more nuanced: that the country’s very complexity makes it the ideal laboratory for real-world AI deployment. More than 120 languages are spoken. Over 90% of the workforce is in the informal sector. A population of 1.4 billion spread in wildly different economic realities.

If AI can be made to work here, the argument goes, it can work anywhere. “The AI models that succeed in India can be deployed globally,” Prime Minister Narendra Modi said in his keynote. “So I invite everyone to design and develop in India and deliver to the world.”

The government put that vision into practice during the summit itself. Agriculture Minister Shivraj Singh Chouhan launched Phase 1 of Bharat-VISTAAR – a multilingual, AI-based conversational platform that allows farmers to access weather data, market prices, crop advisories, pest management information, and government scheme details through phone calls, chatbots, or an app, in Hindi and English to begin with.

For a country where roughly 85% of farmers are small-scale operators, increasingly exposed to climate disruptions and water scarcity, having a farming assistant accessible on a basic mobile phone is not a small idea.

Harder conversations

Amid the summit’s headline announcements, some of the most revealing exchanges were in the sector panels. Government officials overseeing textiles, pharmaceuticals and medical tech announced a study covering more than 350 factories in India to assess how AI adoption can improve quality control, market access, and data flows.

The stakes are substantial: micro and medium enterprises account for 31.1% of India’s GDP and 48.6% of its exports, according to finance ministry data. For these businesses AI is, in theory, the lever that determines whether they remain globally competitive.

“Global competitiveness not depends on labour or wage arbitrage,” said Rohit Kansal, additional secretary at India’s Ministry of Textiles. “AI-driven measurement networks can allow Indian firms to meet global standards at lower compliance and verification costs.” The implication was clear: upgrade, or be out-competed.

The Department of Pharmaceuticals raised the stakes further. Its joint secretary pointed to the deaths of at least 20 children last year from locally produced cough syrups contaminated with an industrial solvent, citing a near-total absence of credible data maintenance at smaller manufacturers.

“If there is no credible data or feedback loop maintained in a particular plant, it is impossible to figure out where the solvent came from. So everything becomes a post mortem,” he said. For an industry where data gaps cost lives, the AI opportunity is also an accountability argument.

The summit’s most concrete proof of concept came from dairy cooperative Amul. Its AI assistant – integrated into a platform already processing milk quality and transaction data from over 3.6 million farmers daily – provides personalised advice in local dialects on everything from herd health to milk quality benchmarks.

“We are getting thousands of calls every day,” said Ajay Sheth, IT manager at Amul. It is a small-scale win, but one that demonstrates the model is at least plausible.

The cracks the summit could not paper over

The India AI summit was not without its frictions. Galgotias University was directed to vacate its stall after presenting a commercially available robot dog manufactured by China’s Unitree Robotics as its own indigenous development – an embarrassment that briefly overshadowed the event’s messaging on Indian AI self-reliance.

Microsoft co-founder Bill Gates withdrew from the summit amid public backlash over his past association with Jeffrey Epstein. And analysts noted that the event’s structure placed multinational corporations on near-equal footing with sovereign governments, while civil society and labour groups had no comparable platform.

The private capital picture remains patchy. “What we’ve not maybe seen as much of right now is venture capital and private equity money to come in to invest in Indian entrepreneurs in the AI space,” Anirudh Suri, founding partner of the India Internet Fund, told CNBC.

The large infrastructure commitments came overwhelmingly from established conglomerates and global hyperscalers. India’s startup ecosystem, by and large, is still waiting for a comparable wave.

Microsoft President Brad Smith offered perhaps the most measured take at the summit, telling CNBC that India could produce its own “DeepSeek moment” in domain-specific AI – an acknowledgement that frugal innovation under constraints can yield breakthrough results. That is the bet India is placing.

Whether it pays off will not be decided in New Delhi. As Kansal put it: “India’s battle for AI will be won or lost on the factory shop floor.”

Want to experience the full spectrum of enterprise technology innovation? Join TechEx in Amsterdam, California, and London. Covering AI, Big Data, Cyber Security, IoT, Digital Transformation, Intelligent Automation, Edge Computing, and Data Centres, TechEx brings together global leaders to share real-world use cases and in-depth insights. Click here for more information.