Индия на пороге ИИ-революции: технологии, данные и пользователи Translation: India on the brink of an AI revolution: technology, data, and users

Sam Altman holds an optimistic view toward India. He remarked that the implementation of AI in the country is «unprecedented globally,» according to The Economist.

India stands as OpenAI’s second-largest market in terms of users and may soon become the top one. In August, the company introduced a local subscription priced at just $4.6 per month, making it the most affordable option. In contrast, the cheapest plan available elsewhere costs $20.

Later in 2025, the firm intends to establish an office in New Delhi, and Altman plans to visit India at the end of September. Reportedly, during his visit, he will present a proposal for creating a data center within the country.

Other tech giants are also showing interest in the nation:

This «competition» offers Indians access to cutting-edge solutions at low costs. For AI startups, the advantage lies in reaching and securing hundreds of millions of users and the data they generate.

The opportunities are immense. With roughly 900 million internet users, only China has a greater number. However, unlike China, India is open to American tech companies.

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

However, AI startups may find it challenging to derive significant profits from this market. Most tech companies offer lower prices locally. For example, a Netflix subscription in India costs just $1.69 per month compared to $7.99 in the USA.

For cloud services with minimal expenses, this isn’t critical, but fulfilling AI requests can be costly. Data processing for an average user costs about $0.07 per million tokens, with a single response potentially consuming hundreds or thousands of tokens. These expenses are similar whether in Bangalore or Silicon Valley.

Requests to ChatGPT utilize roughly ten times the computational power compared to Google. Generating a single image with AI requires the same amount of electricity as fully charging a smartphone.

Beyond energy, data centers need significant amounts of water for cooling — about 500 milliliters are needed for every 5-50 inquiries to ChatGPT.

Dmitry Shevelenko, the commercial director of Perplexity, acknowledged that offering the service for free for a year is expensive, but the company has time to «establish itself» and convert users into paying customers.

He noted that India displays one of the highest engagement levels among the countries where Perplexity has conducted tests, predicting it will become an «attractive subscription market» within five years.

Subscriptions are not the only incentive. India possesses a vast, diverse, and chaotic base of internet users, covering individuals from numerous linguistic backgrounds, making it an ideal «testing ground» for new products and large-scale user behavior monitoring.

Companies have begun to notice that Indian users prefer speaking to AI models rather than typing queries, likely due to low literacy rates among the population — many are unable to read or write, reports The Economist.

AI startups have mainly exhausted all significant public datasets available for training and improving models. Consequently, Indian users have emerged as a valuable new source of data.

The country has established a robust digital infrastructure known as India Stack, integrating government services — a biometric identification system and digital payments. This has led to hundreds of millions of individuals getting online. Their queries aid in precisely tuning neural networks.

India’s regulatory environment also supports this. Conark Bhandari from the Carnegie Endowment for International Peace pointed out that regulations do not hinder companies from transferring data abroad.

«There’s nothing in the legislation preventing the use of information collected in India to train models overseas,» he added.

Indian users welcome the entry of foreign AI firms, although some express concerns over the impact on local players and a long-term dependence on American platforms.

Anxieties heightened after former President Donald Trump imposed high tariffs on goods from India. Venugopal Garre from Bernstein believes that American companies possess greater resources and infrastructure, which could «kill India’s prospects» in AI by deterring investments in local startups.

Altman, however, presents a more hopeful perspective. He asserts that the country could become «one of the leaders of the AI revolution.» The only question remains about what kind of leader it aims to be: one that dominates through its vast user base or one that creates its own technologies?

Additionally, the India-based startup QpiAI, focused on integrating artificial intelligence with quantum computing, secured $32 million in a new Series A funding round, with the Indian government as a co-investor.