India: The Next Catalyst for the AI Boom

Sam Altman is optimistic about India's AI potential.

Sam Altman is optimistic about India. He states that the implementation of AI in the country is “unparalleled in the world,” reports The Economist.

The country is the second-largest market for OpenAI by user numbers and may soon become the first. In August, the company launched a local subscription at the lowest price of $4.6 per month. Elsewhere, the most affordable plan costs $20.

Later in 2025, the firm plans to open an office in New Delhi, and at the end of September, Altman intends to visit India. According to media reports, during his visit, he will present a concept for establishing a data center in the country.

Other tech giants are also interested in the country:

• In January, Microsoft pledged to invest $3 billion in expanding local AI infrastructure;
• In August, Google and Meta announced a partnership with Reliance Industries — a conglomerate owned by India’s richest man, Mukesh Ambani — to build data centers and integrate artificial intelligence into local businesses;
• In July, the startup Perplexity signed an agreement with Bharti Airtel to provide its AI service free for a year to all 360 million customers of one of the largest telecommunications companies.

For Indians, this “battle” promises access to cutting-edge solutions at a low cost. For AI startups, the benefit lies in reaching and potentially securing hundreds of millions of users and the data streams they generate.

Inexhaustible Resources

The opportunities are immense. The country has about 900 million internet users — second only to China. However, unlike China, the country is open to American tech companies.

Google’s Android operates on more than 90% of smartphones in the country. Meta-owned WhatsApp has over 500 million active users. Indian e-commerce is also dominated by American giants: Amazon and Flipkart (Walmart).

AI’s Voracious Appetite

However, it will be challenging for AI startups to make substantial profits from this market. Most tech companies offer lower prices locally. For instance, a Netflix subscription in India costs only $1.69 per month compared to $7.99 in the US.

For cloud services with low overheads, this is not critical, but executing AI queries is costly. Data processing for a standard user costs about $0.07 per million tokens. Answering a single question can involve hundreds or thousands of tokens. These costs are the same whether in Bangalore or Silicon Valley.

Queries to ChatGPT require about 10 times more computing power compared to Google. Generating a single AI image consumes as much electricity as fully charging a smartphone.

Besides energy, data centers consume a lot of water for cooling equipment — 500 milliliters are used for every 5-50 questions to ChatGPT.

Perplexity’s Chief Commercial Officer Dmitry Shevelenko acknowledged that providing the service for free for a year is costly, but the company has time to “establish itself” and convert users into paying customers.

He noted that India shows one of the highest engagement levels among countries where Perplexity has conducted tests. In five years, it will become an “attractive subscription market.”

Beyond Money

Subscriptions are not the only incentive. India’s internet user base is vast, diverse, and chaotic. It spans people from numerous linguistic groups and serves as an ideal “testing ground” for new products and observing user behavior on a large scale.

Companies have begun to notice that Indian users prefer talking to AI models rather than typing queries. This is likely due to the low literacy rate — many cannot read or write, writes The Economist.

The Data Dilemma

AI startups have mostly exhausted all major public datasets available for training and improving models. Thus, Indian users have become a valuable new source of information.

The country has established a robust digital infrastructure called India Stack, which integrates government services — a biometric identification system and digital payments. This has brought hundreds of millions of people online. Their queries help fine-tune neural networks more accurately.

India’s regulatory regime also supports this. Konark Bhandari from the Carnegie Endowment for International Peace noted that the rules do not restrict companies from transferring data abroad.

“There is nothing in the legislation that prevents using information collected in India to train models overseas,” he added.

Indian users welcome the arrival of foreign AI companies. However, some are concerned about the impact on local players and long-term dependence on American platforms.

Concerns have intensified following US President Donald Trump’s imposition of high tariffs on goods from India. Venugopal Garre from the research firm Bernstein believes that American companies, with their vast resources and powerful infrastructure, could “kill India’s prospects” in AI by deterring investment in local startups.

Altman offers a more optimistic vision. He suggests that the country could become “a leader in the AI revolution.” The question is what kind of leader it chooses to be: one that dominates through its massive user base or one that creates its own technologies?

Back in April, the Indian startup QpiAI, focused on the joint application of artificial intelligence and quantum computing, raised $32 million in a new Series A round. The Indian government acted as a co-investor.