Why Infosys’s share buyback highlights India’s big AI vacuum

However, there are a couple of questions to consider: Should Indians be concerned about the scarcity of AI ventures in India? Why is Infosys using its funds to repurchase its shares rather than invest in developing AI tools?

Michael Burry, an American investor immortalised in the film The Big Short, persuaded Goldman Sachs and several other investment banks to create and sell to his hedge fund, Scion Capital, credit default swaps on tranches of subprime mortgages assembled into synthetic instruments called collateralised debt obligations (CDOs).

Few understood these securities, but everyone was willing to invest in them because rating agencies had rated them as more than investment-grade. That was in 2005, as the US housing bubble was building up, fuelled by subprime loans.

When the bubble burst, Burry and his hedge fund made a killing. It more than quadrupled its initial outlay because he stuck with his conviction, despite pressure from investors to cash out.

Earlier this month, Burry bought put options worth more than $9 million in two AI heavyweights, Nvidia and Palantir.

Nvidia, which makes processors used to generate AI, has a 12-month trailing price-to-earnings multiple of around 54. Palantir, which utilises AI to provide solutions to companies and governments, has a price-to-sales ratio of 86 and a forward price-to-earnings ratio of nearly 250.

Following Burry’s bet, Palantir’s shares slid about 8%, despite a big jump in third-quarter earnings. Nvidia’s share prices also fell, but later recovered on solid earnings and growth outlook.

Cross-interests among tech companies riding the AI boom ensure that if confidence cracks in one player, all could fall together.

In October, Nvidia announced a partnership with Palantir and others, while OpenAI acquired a warrant to own a stake in chipmaker AMD.

India vacuum

While several Indian IT services companies have started mentioning AI as part of their services, there are no mainstream, AI-focused domestic companies that have experienced a boom or a bust. But, is it a great feeling to have?

While developing a large language model (LLM) from scratch is a tall order for any company, developing applications using an open-source or licensed model from any of the established players is less challenging, but probably a more impactful activity.

West Asian countries are trying to leverage their energy abundance to become major players in data centres and develop AI based on the capacity of these data centres.

Saudi Arabia is investing sovereign wealth into HUMAIN—a massive programme spanning data centres, foundation models, local-language AI and industrial AI solutions, especially for the oil and gas industry. The UAE has a full-fledged National AI Strategy 2031.

Meanwhile, China is racing ahead to deploy AI in every facet of doing business and gain a competitive advantage.

But what are Indian companies doing? How do we see Infosys using its funds to buy back shares, rather than create new AI-focused businesses?

What India could do

While the government can pioneer development of an indigenous foundation model, individual companies can focus on developing applications and business solutions based on open-source or licensed AI models.

Like IBM, Indian companies can offer enterprise-wide data intelligence that goes beyond the sectoral solutions companies tend to experiment with.

India has linguistically homogenous regions as large as independent nations of the world. AI models that run in Indian languages would boost the productivity of small and micro enterprises, which already use digital payments and GST systems.

It may feel safer to sit out a potential bubble, but it is riskier not to participate at all. It is better to be part of a bubble that eventually deflates but leaves behind long-term survivors than to avoid the froth entirely and watch domestic companies fall behind global rivals.