OpenAI’s Big Edtech Plan For India Has A Data Localisation Problem

Picture this: Shaig Abduragimov, a solutions engineer at OpenAI, flying from London to New Delhi to present a product he has been building. As he speaks, a South Delhi school principal is frantically taking notes.

This was OpenAI’s formal entry into India’s edtech sector. The AI giant, which is preparing to open its New Delhi office, invited leading institutions, universities, edtech startup founders, and well-known educators such as Anand Kumar of Super 30 to a small gathering in the capital.

The event, titled “Education Summit, India” showed off OpenAI’s plans for products and the leadership for the Indian education vertical. 

Leah Belsky, OpenAI’s Vice-President of Education and a former Coursera managing director, will lead the company’s education strategy for India and the wider Asia-Pacific region. 

Primarily, the company is aiming to partner with schools and colleges to offer a copilot-like experience for students and teachers. 

So why is OpenAI pushing into Indian classrooms, and what kind of revenue potential does it see?

OpenAI’s Edtech Endgame

According to Pragya Misra, OpenAI’s head of policy and partnerships in India, the aim is not financial. “The aim is not to make revenue,” Misra told Inc42. “The company just wants to enable education in India and is not looking for revenue opportunities.”

There is some credibility to this given OpenAI’s complex corporate structure. The organisation is formally a non-profit, but it controls a for-profit Public Benefit Corporation (PBC) responsible for raising capital and developing AI models. The latter is bound to pursue the non-profit’s mission and is subject to capped profits.

Yet the commercial reality is harder to ignore. Microsoft, OpenAI’s largest investor, holds exclusive rights to resell OpenAI’s APIs through its Azure cloud platform. While OpenAI develops AI and promotes its societal benefits, Microsoft actively sells the tools to customers worldwide, and will continue to do so.

Belsky noted that a large share of early users testing OpenAI’s Study Mode came from India, and that their feedback had been crucial in improving the platform. Study Mode was launched only in July, but Microsoft had already been working with Indian edtech giant PhysicsWallah (PW) to trial similar tools well before that.

In February, Microsoft announced that PW had launched Alakh AI, a learning platform powered by OpenAI’s GPT-4o model through Azure. The suite included an AI Guru for teaching concepts, a Smart Doubt Engine for instant answers, an AI Grader for marking assignments, and Sahayak, a digital assistant to support students and teachers.

However, Alakh AI was actually introduced in December 2023, meaning that OpenAI has been working not only with PW but also, in all likelihood, with other for-profit edtechs in India for at least two years. 

At the same time, it has also collaborated with nonprofits such as Khan Academy in early 2023.This mix of alliances makes it difficult to accept the claim that OpenAI’s role in India is purely about “investing in the ecosystem”. 

The tools it has built so far appear better suited to private edtech companies and fee-paying institutions than to the broader public sector.

And even if it wants to, will the government agree? The same government which has set apart an outlay of INR 10,372 Cr for five years for the India AI Mission, to provide an alternative to global LLMs. 

Can OpenAI Answer The Localisation Question?

But going beyond private institutions and edtech companies will be a tough ask for OpenAI. At the moment, the likes of IIT, IIMs are out of reach for the company. 

The government is unlikely to welcome OpenAI with open arms unless the company bends to its rules, at the centre of which are questions around data localisation and data sharing. 

On the sidelines of OpenAI’s Education Summit, IndiaAI Mission CEO Abhishek Singh clearly stated what is expected of AI companies looking to build tools for education. In short, it’s about open-source models, strict data localisation, and tools built for mass public use, not just elite schools and edtech platforms.

Singh questioned OpenAI’s closed approach, even quipping that “the open is in your name itself. So, it’s high time OpenAI becomes open now.” 

The government wants models that can be hosted on government infrastructure, made open-source, and shared widely for public benefit.

Data sovereignty is another sticking point. India has become increasingly protective of where its citizens’ data is stored. 

Singh highlighted that the government’s policy is to keep public-use AI models inside India’s borders. That runs directly against OpenAI’s current setup, where data typically flows through Microsoft’s Azure cloud servers abroad.

At the same time, the government is looking at AI as a mass tutor for underserved regions. Singh gave the example of a village in Nagaland where no science or maths teachers are available. In his vision, an AI tutor, speaking local languages and hosted on government platforms, could fill that gap. 

This is not the market OpenAI usually caters to, but it is the one the Indian government cares about. And at the moment, OpenAI is not in the loop when it comes to these solutions.

While OpenAI thrives on a closed, profit-capped structure tied deeply to Microsoft, India wants open-source, low-cost tools that can run locally and affordably. 

Why India, And Why Now?

OpenAI’s move into India is not accidental, nor is it purely altruistic as the company might want to position it as. India represents a perfect balance of scale, data, and commercial upside that OpenAI cannot afford to miss. 

Most recently, OpenAI rolled out its lowest pricing model for ChatGPT in India at INR 399 a month, and allowed users to pay for subscriptions via UPI, another strategy to capture the digitally native young Indian consumer market. 

Beyond the consumer base and in the education sector, India offers OpenAI three things it cannot get anywhere else at this scale: users (students and educators), data, and a massive TAM. With more than 250 Mn students in schools and higher education, India is the world’s largest education market outside China.

For an AI model built to learn and adapt, this is a goldmine of real-world usage. OpenAI’s Belsky admitted that Indian students already make up a significant share of those using Study Mode, feeding the platform with invaluable data to refine its tools.

Beyond data, India is also a proving ground. For OpenAI, that means a chance to trial products with millions of users, pressure-test them at scale, and then carry polished versions to higher-paying Western markets. India, in effect, becomes the laboratory.

The commercial incentive, however, lies one step removed via Microsoft, which can turn every Indian edtech tie-up into a revenue stream. The abovementioned partnership with PW may represent a playbook: OpenAI provides the underlying model, an edtech company/ university brings the students, and Microsoft sells the infrastructure. 

For Azure, this is less about altruism and more about securing a long-term pipeline of enterprise customers from one of the fastest-growing edtech sectors in the world.

Then there’s the timing. India is in the middle of an AI land grab. Google is pushing Gemini into schools, homegrown edtechs are experimenting with AI tutors, and the government itself is promoting AI-led education initiatives. Perplexity’s free premium subscription is another sign that the frontrunners in the AI revolution are now eyeing India more seriously. 

And now with an office in Delhi, OpenAI is looking to ensure it is not crowded out by rivals.

But the bigger question remains: will OpenAI’s India play truly serve students and educators, or will it primarily fuel Microsoft’s commercial ambitions? And will the government’s emphasis on sovereign AI scupper these big plans before they fully fledge?