Indian firm seeks $3 billion to buy Nvidia next-gen chips

A massive infrastructure push

AM Intelligence, a company based in Hyderabad, is moving to secure significant funding for its AI ambitions. According to a report from Bloomberg, the firm is looking to raise up to $3 billion through a combination of bank loans and bond issues. These funds are specifically earmarked for purchasing Nvidia's next-generation Rubin graphics processing units (GPUs).

The company describes itself as one of the largest committed buyers of these specific chips outside of North America. By securing these high-end components, the firm intends to build a robust foundation for large-scale AI operations. This strategy positions the company to provide essential hardware for the growing demand of artificial intelligence applications.

The financial plan also includes a projected $300 million in advance payments from potential offtakers. This suggests that there is already significant interest from clients who want to lease or use the computing power once it is online. The firm is leveraging these early commitments to bolster its debt-raising efforts in a competitive market.

Scaling across borders

The company's broader investment plan involves spending $6 billion to establish 100 MW of computing power. This capacity will be split across data centers located in both India and Malaysia. This geographic spread allows the firm to tap into different markets while building a diversified infrastructure network.

To reach this 100 MW goal, AM Intelligence plans to purchase a total of 29,000 Nvidia Rubin chips. These chips are specifically designed for agentic AI, which refers to systems capable of performing complex tasks autonomously. By focusing on these specific units, the firm aims to gain a first-mover advantage in the Indian market.

Beyond the immediate 100 MW goal, the company has even more ambitious targets for the near future. It plans to add another 300 MW of capacity to the market within the next 15 months. This expansion will require an investment of more than $20 billion, highlighting the massive scale of the current AI infrastructure build-out.

The rise of neoclouds

AM Intelligence is part of a broader trend where Indian energy companies are transforming into what experts call neocloud companies. These firms are setting up data centers with massive computing power that can be leased to hyperscalers, which are large-scale cloud service providers. This shift allows energy companies to diversify their business models by providing the backbone for the digital economy.

Several major Indian entities are already participating in this transition, including the Adani Group, Reliance, and Tata Power. These companies are leveraging their existing energy expertise to support the high electricity demands of AI data centers. By providing both the power and the computing hardware, they can offer a comprehensive service to tech giants.

This trend is supported by the growing need for localized infrastructure. As AI becomes more integrated into business operations, companies want to ensure that the underlying hardware is reliable and accessible. The move toward computing-as-a-service allows smaller enterprises to access high-end chips without having to own the physical hardware themselves.

From training to agents

The focus of the AI industry is expected to shift significantly in the coming year. While current efforts often center on training foundational models, the next phase will likely prioritize running production agents. These agents are designed to perform specific, low-latency tasks for businesses rather than just learning general information.

Nvidia's Rubin GPUs are particularly well-suited for these types of agentic workloads. Because they are optimized for speed and specific actions, they are becoming a primary target for infrastructure providers. AM Intelligence is positioning its hardware to meet this specific demand as enterprises move toward practical AI deployment.

The company is also developing a much larger long-term roadmap that includes 5 GW of AI data centers. This plan covers regions including India, the United States, and Europe. By expanding globally, AM Intelligence aims to become a major international provider of compute-as-a-service over the next several years.