AM Intelligence Seeks $3 Billion For Nvidia Rubin Chips
The firm aims to secure new compute capacity for its global network of AI data centers within two years.
Updated on Oct. 30, 2026 in Data Centers
Live Poll
Is now a good time for firms to invest heavily in expanding AI computing infrastructure?
Hyderabad-based firm AM Intelligence is seeking to raise $3 billion in debt to finance the purchase of 29,000 Nvidia Rubin chips. The company plans to deploy these processors as part of a $6 billion investment to establish 100 MW of initial computing power across facilities in Malaysia and India.
Why it matters
This debt-raising effort signals an aggressive scaling of AI infrastructure capacity by a private entrant, aiming to secure high-demand hardware amid a global scramble for specialized compute resources. The firm is betting on a rapid expansion of data centers to meet the growing compute-as-a-service market.
The firm plans to integrate 29,000 Nvidia Rubin units, a GPU architecture displayed in October 2026, to reach an initial 100 MW capacity. Future scaling targets include 400 MW of capacity within two years and a broader long-term goal of 5 GW in AI data centers.
The players
AM Intelligence
A Hyderabad-based firm focused on the development of AI data centers and compute-as-a-service infrastructure.
Nvidia
The designer of the Rubin graphics processing unit architecture used in high-performance AI compute clusters.
Greenko
An energy company whose founders own AM Intelligence and are shifting focus toward large-scale AI compute infrastructure.
The details
AM Intelligence, founded by the owners of Greenko, aims to build out compute-as-a-service capacity by leveraging capital from bond issues and bank loans. The strategy involves securing $300 million in advance payments from potential customers to partially fund the heavy hardware acquisition costs associated with Nvidia's Rubin graphics processing units — specialized chips designed for high-throughput AI model training and inference. The company is managing a geographically distributed portfolio of centers across India, Malaysia, the U.S., and Europe.
Timeline
October 5, 2026: AM Intelligence announced a $6 billion investment plan.
October 21, 2026: Nvidia displayed Rubin computer modules in Berlin.
October 30, 2026: The $3 billion debt-raising strategy was reported.
Next 15 months: The firm intends to add 300 MW of capacity.
Within two years: The company targets 400 MW of total compute capacity.
The Tech Race
AM Intelligence is positioning itself alongside established cloud providers by aggressively securing the latest Nvidia hardware for its compute-as-a-service offerings. This race to deploy Rubin-based centers mirrors wider industry attempts to lock in specialized hardware supply chains for next-generation AI workloads.
The firm aims to bring 300 MW of new capacity to the market within 15 months through a $20 billion investment phase. Enterprises looking for compute-as-a-service will monitor these facilities as they move from funding to full operational status in India and Malaysia.
The takeaway
The company is betting its future on the rapid deployment of Nvidia's newest hardware at a massive scale. Observers should track the realization of the planned 300 MW capacity expansion to see if the firm meets its aggressive 15-month timeline for project delivery.
Further reading
For more on the infrastructure shifts defining this sector, visit the Data Centers section.
Source note: This article includes information reported by CNBC.
Live Poll
Is now a good time for firms to invest heavily in expanding AI computing infrastructure?
