This week’s AI Impact Summit in New Delhi, India made a noticeable splash in the world of AI. Attended by 300K people and most of the CEOs of the leading global AI firms, the summit was the 4th in an annual series previously held in the UK (Bletchley Park), South Korea (Seoul), and France (Paris), respectively. On stage with Prime Minister Narendra Modi were Google DeepMind CEO Demis Hassabis, Google CEO Sundar Pichai, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Microsoft President Brad Smith, and Meta’s Chief AI Officer Alexandr Wang, among others. India has been vocal about its desire to become a global AI hub, aiming to attract $200B+ in AI infrastructure investment by 2028 (roughly on par with what the EU has committed for AI). At the summit, global CEOs vied for the spotlight with major announcements about their planned investments in India.

India has emerged as one of the most attractive markets for AI. It is the largest country by population with 1.5B people (overtaking China in 2023), the youngest large country with a median age of 28 (vs. 39 in US and 41 in China), and one of the most connected with 1B broadband subscribers. 62% of Indians reportedly use generative AI tools (more than any other country except Brazil). Indians share 33% more data with AI tools than US users – one reason why India produces 20% of the world’s data. OpenAI has 100M weekly active ChatGPT users in India (of its 700M+ total), with 50% in the 18-24 age bracket (vs. 33% globally). Both Anthropic and Replit have said India is their 2nd-largest market in users, after the US. For Google, India is its #1 market in usage of Gemini for learning and in image-generation usage. The adoption of AI in India, in large part, has been bolstered by the introduction of sub-$5 subscriptions (e.g. OpenAI’s ChatGPT Go, Google’s AI Plus) and free versions made available through mobile carriers.

The AI Impact Summit builds upon the Indian government’s IndiaAI Mission (Mar 2024). The government allocated $1B+ towards building a sovereign AI ecosystem, including compute, AI research, datasets, applications, startup financing, reskilling, and AI safety. As of 2024, India was ranked #3 in Global AI Vibrancy by Stanford’s Institute for Human-Centered AI, after the US and China.

Data centers are a major pillar of the IndiaAI Mission, and key to both AI and local data sovereignty for India. India houses only 3% of the world’s data, despite producing 20% of it. India is now adding 20K GPUs to the existing 38K GPUs of shared compute capacity under the IndiaAI Mission, and aiming for 100K+ GPUs to be online by end of 2026. India is currently proposing to offer global big tech firms a deal for zero taxes until 2047 on revenues from all cloud services that are run from data centers inside India and sold outside of India.

Some of the biggest commitments announced this past week were data-center investments by Indian giants. Infrastructure and energy conglomerate Adani Group pledged $100B for large renewables-powered data centers by 2035, with focus on supporting Indian LLMs (large language models) and data sovereignty. It is partnering with Google and Walmart-backed Flipkart on AI data centers, and in talks for more large-scale campuses, for a projected total of 5 GW of capacity. Adani expects the $100B investment will trigger another $150B in investment in related industries, for a total impact of $250B over the next decade. Digital and energy conglomerate Reliance Industries followed with a $110B commitment over 7 years for multi-gigawatt AI data centers and services, with one 120-MW data center expected to come online in H2 2026.

US cloud and AI players were not shy about announcing their own commitments. At the summit, Google CEO Sundar Pichai revealed a new America-India Connect initiative to lay new fiber-optic lines between India and the US/other countries; a DeepMind partnership with the Indian government on frontier models; and partnerships in education, agriculture, and energy. Google is all in on India, with its largest campus outside the US set to open in Hyderabad this year, a $15B data center (announced Oct 2025) being built in Andhra Pradesh, and partnerships with Reliance and Adani. Microsoft is opening a new cloud region in Hyderabad this year and investing $17.5B in cloud and AI over 4 years. Similarly, Amazon is investing $35B+ across its businesses in India thru 2030.

Among the chip players, Nvidia is partnering with Indian cloud providers Yotta, Larsen & Toubro (L&T), and E2E Networks to provide AI chip clusters for new data centers. AMD is partnering with Tata Group’s TCS to bring AMD’s Helios rack-scale AI architecture to Indian data centers. India has also been making a push over the past 4 years to foster its own homegrown chip industry, with $18B in chipmaking projects already underway.

Pure-play AI companies OpenAI and Anthropic are opening new offices and leaning heavily on partnerships. OpenAI is opening offices in Mumbai and Bengaluru, in addition to its existing Delhi office. OpenAI has tie-ups with Tata Group on data centers and AI services; with Reliance to add AI search to streaming service JioHotstar; with fintech Pine Labs to integrate AI into its payment processes; with higher-education institutions (e.g. IIT Delhi, IIM Ahmedabad) to integrate AI into core academic workflows; and other partnerships with Eternal, Cars24, HCLTech, PhonePe, CRED, and MakeMyTrip. Rival Anthropic is launching a new office in Bengaluru (its second in Asia, after Tokyo). Anthropic is partnering with Infosys to build AI agents, and with an array of Indian organizations on business solutions, education, and public-sector use cases.

India’s startup ecosystem had some encouraging news this month, after a relatively slow 2025 and a subdued IPO market so far in 2026. The government approved a $1.1B state-backed venture fund-of-funds, and allowed startups of up to 20 years old and $33M in revenue to be eligible for startup-specific tax, grant, and regulatory benefits. US VC General Catalyst this week announced plans to invest $5B in India over 5 years (up from $500M-$1B before). Other large VC firms are expected to follow suit with similar announcements. Nvidia revealed collaborations with VC firms (e.g. Activate, Peak XV, Z47, Elevation Capital, Nexus Venture Partners, Accel India) to identify, build relationships with, and fund Indian startups. (Nvidia already has 4,000+ Indian AI startups in its Inception Program.) Indian startups have had some prominent successes lately – with vibe-coding platform Emergent revealing $100M+ in run-rate revenue after launching just 8 months ago, AI cloud startup Neysa securing $600M in equity funding, and power-delivery startup C2i raising a $15M Series A.

Deeptech startups have been a particular area of focus for India, with the introduction of the India Deep Tech Alliance of VC/PE firms (Sep 2025) and affiliated UChicago-led India Deep Tech Accelerator (Feb 2026). The Indian government is aiming to have 50+ deeptech startups emerge from India over the next few years.

While the talent in India will need to be reskilled, there is a large base of IT talent to be redeployed. The push in India comes at a time when Indian IT services have faced a steep selloff (mirroring US SaaS companies), with India’s Nifty IT index down 26% since Jan 2025 and down 15.5% in just 2026 YTD. India’s $283B IT services industry employs 5.8M workers and contributes 7%+ to national GDP (as of FY25). The ex-CEO of IT services giant HCL warned recently that IT services companies are likely to be focused on profits rather than jobs. The momentum appears to be shifting away from IT services and towards global capability centers (GCCs) set up by multinationals.

As one analyst put it, “If there is a single domain where India’s AI ambitions will succeed or fail, it is energy.” India’s energy infrastructure can be notoriously unreliable and plagued with market distortions and financially stressed power-distribution companies. On the other hand, more than half (51%) of India’s energy already comes from renewable sources. Partnerships with energy conglomerates like Adani and Reliance are key to ensuring that data centers have access to captive power that is reliable and economically viable.

While India has big AI ambitions, people are describing its approach as a “frugal AI strategy” that could provide a blueprint for countries that can’t afford US-style AI infrastructure. Many of the efforts underway have placed emphasis on being localized, multilingual, efficient (e.g. smaller, shared commons), and accessible (e.g. usable offline). India has “22 constitutionally recognized languages and over 1,500 more recorded by the country’s census.” AI players like Google, Anthropic, and Nvidia are tailoring their model training and products with an eye towards India’s most spoken languages. Cohere just unveiled a new family of open-weight multilingual models called Tiny Aya, which support 70+ languages and can run on a standard laptop without internet. In the same vein, Indian AI lab Sarvam and government-backed sovereign-AI consortium BharatGen both introduced new models available in 22 Indian languages, trained from scratch on Indian data using IndiaAI MIssion compute resources. Like the Cohere models, Sarvam’s open-source models are designed to be smaller, efficient models that can be used on edge devices.

Aug 9 2024 (3 Shifts): Generative AI threatens India’s BPO industry

Jan 30 2020 (Brief #21): India is the market battleground everyone is watching