img

Andrew Feldman, the CEO of AI chip maker Cerebras System, has stressed the increasing capacity for AI leadership from countries like India, pointing out that world-class models can now be built by small, skilled teams without needing much capital.

In his interview with Moneycontrol at the Web Summit Qatar 2025, Feldman cited the recent achievement of DeepSeek, an AI foundation model that was developed by a small team, as proof that the competition in AI no longer requires billions in investment.

Those skilled in lesser developed regions, such as India, can also compete due to lowered costs of capital, Feldman told Moneycontrol.

The achievement of DeepSeek sparked new discussion around India’s AI goals and proved that countries with adequate talent and framework can create competitive AI systems without the need for substantial funds.

When asked about the focus of India in developing large foundational models/sub-models that are specific in terms of the domain, Feldman pointed out that the focus is not on scale but the intent behind the model, is what matters.

“The debate regarding macro and micro models should lie in the case metric and the data,” he says. “That is where companies ought to concentrate on.”

Large models serve and are core necessities for profoundly sophisticated reasoning and broad tasks; however, smaller models are utilized in many instances as they are more cost-effective and efficient.

The Silicon Valley AI chip firm Cerebras Systems has also stepped up its investment in India with data centers and talent hiring.

The company, which specializes with wafer scale AI processors, is concerned with India because of its renowned universities and large pool of skilled professionals.

“India must establish AI and it should be run on Cerebras systems. That’s why we want to step our investment in India gradually. We have an office in banglore as well some people in hyderabad, and we hope to expand,” Feldman said.


Read More: Samsung Galaxy S25 Plus Review: The Perfect Middle Ground in the Flagship Lineup