India is preparing for a major AI computing expansion. AM Intelligence, backed by the people behind Greenko, has ordered another 20,000 NVIDIA Rubin GPUs. The scale of this move puts Hyderabad at the centre of a competition for infrastructure that could change how India handles AI compute power.
AM Intelligence has ordered about 20,000 NVIDIA Vera Rubin GPUs. This is part of a plan to expand its AI infrastructure in India and Malaysia. The new order is set to be delivered in the quarter of 2027. This brings the number of Rubin GPUs that AM Intelligence has committed to around 29,000.
This news matters because AI development is becoming a race not for better models, but for access to the computing power needed to train and run them. AM Intelligence is trying to provide compute-as-a-service, aiming to help big companies, AI businesses, enterprises, developers, and government AI programs.
This new commitment follows AM Intelligence’s August 2026 announcement that it would get 9,000 NVIDIA Rubin GPUs for its AI factory in Hyderabad.
The new orders will add around 70 MW of capacity in India and Malaysia. It brings AM Intelligence’s committed capacity to almost 100 MW. The systems will be set up as NVIDIA Vera Rubin NVL72 rack-scale platforms. These are supported by networking, advanced storage and liquid cooling systems.
For Hyderabad, this development is very special. The city is already a centre for technology and data centres. Now the AI factory gives the city a role in the infrastructure needed to support the wave of AI workloads.
This isn’t about an AI app or a funded startup. The investment is in the physical computing infrastructure needed to support these services.
Also Read: India’s 4G Phones Are Making a Comeback
The company has more plans
AM Intelligence says the order for 20,000 GPUs is only part of a much bigger expansion plan.
The company plans to add another 300 MW of AI compute capacity to the market in the next 15 months. This will involve more than $20 billion in money spent. This is in addition to the $6 billion already set aside for the 100 MW.
The company’s global Compute-as-a-Service plan is currently 400 MW. This includes India, the United States, Europe and Malaysia.
The long-term goal is even bigger. AM Intelligence wants to reach 1 GW of Compute-as-a-Service capacity. It also plans to build around 5 GW of AI data-centre capacity in India, the US and Europe.
Why NVIDIA Rubin is important
The infrastructure will use NVIDIA’s Vera Rubin platform. These are NVL72 rack-scale systems made for AI tasks.
AI models need massive computing power to train and run. As companies move from trying AI tools to using AI agents, reasoning models, and big workloads in life, the need for powerful GPUs is growing.
AM Intelligence bets that businesses will want to rent AI computing power rather than build their own large GPU systems.
The company’s idea is to make powerful computing available as a service. This lets AI developers and companies access large infrastructure without building their own AI data centres.
Also Read: Apple, Fitbit, Whoop, Pebble: Fitness Trackers Compared
India’s AI infrastructure competition is speeding up
This announcement also shows a change in India’s AI goals.
India has been trying to improve its AI ecosystem through computing projects, data centres, chip programs, and cloud setups. Building strong AI services requires significant physical computing power.
The AM Intelligence investment shows that private companies are now making bets like this.
For Indian AI companies, the big issue is not just whether they can build. It is whether they can get computing power to train, adjust, and put that model to use at a good price.
With 29,000 Rubin GPUs already promised and a much bigger plan in the works, AM Intelligence is positioning itself as a key infrastructure provider to meet that need.
The big news is 20,000 GPUs. The bigger story is that India is becoming a hub for the physical infrastructure that supports the worldwide AI boom.
Image Credits: (X- nishant568)
If this article saved you time or helped you decide better, consider supporting our work.
FAQs
How many NVIDIA GPUs has AM Intelligence ordered?
The company has placed binding orders for around 20,000 additional NVIDIA Rubin GPUs.
How many GPUs has AM Intelligence committed in total?
The latest orders take its total committed Rubin GPU capacity to around 29,000.
Where will the new GPUs be deployed?
The additional capacity will be deployed across India and Malaysia, with deliveries scheduled for Q2 2027.
What role does Hyderabad play?
AM Intelligence’s first AI factory is being developed in Hyderabad and includes its initial commitment of 9,000 Rubin GPUs.
How big is AM Intelligence’s long-term plan?
The company is targeting 1 GW of Compute-as-a-Service capacity and 5 GW of powered AI data-centre capacity across multiple markets.
Support Telecomtalk
