Government jobs • Exam updates • PreparationIndependent information portal
RojgarRankYOUR NEXT STEP STARTS HERE
Current Affairs

AM Intelligence Orders 20,000 More Nvidia Rubin GPUs for India and Malaysia

By Published

AM Intelligence announced on 5 October 2026 two more binding orders for 20,000 Nvidia Rubin GPUs, with delivery scheduled for Q2 2027. This explainer covers the facts, the gaps in the evidence and the exam-relevant terms.

On this page
AM Intelligence Orders 20,000 More Nvidia Rubin GPUs for India and Malaysia — title card
RojgarRank title card.

AM Intelligence (AMI), an AI infrastructure company set up by the promoters of Greenko, announced on 5 October 2026 that it has placed two more firm and binding orders for 20,000 Nvidia Rubin GPUs. The GPUs are for sites in India and Malaysia, and delivery is scheduled for the second quarter of 2027. As of 6 October 2026 the orders are announced, but the GPUs have not been delivered and the new capacity is not yet running.

AM Intelligence's Nvidia GPU order: key facts

What was announced and when

AMI's press release is dated Hyderabad, 5 October 2026. GKToday carried the story on 6 October 2026, so the event date and the report date differ by a day. Details below come from these two sources.

ItemWhat the sources sayStatus as of 6 Oct 2026
First order9,000 Nvidia Rubin GPUs for AMI's first AI factory in Hyderabad, announced 25 August 2026Announced; delivery date not given in the evidence
New ordersTwo further binding orders for 20,000 Rubin GPUs in total, about 70 MW, for India and MalaysiaAnnounced 5 October 2026
Delivery of new GPUsQ2 2027Scheduled, not completed
Total committedAbout 29,000 Rubin GPUs and close to 100 MW across the three ordersCommitments, not installed capacity

The technology involved

The GPUs will be deployed as Nvidia Vera Rubin NVL72 rack-scale systems. AMI's chairman says the sites will pair the Vera Rubin platform with RoCE networking (RDMA over Converged Ethernet) and advanced storage. Each site is described as built for high rack power density and liquid cooling, so that later generations of AI chips can fit in the same footprint.

Why the order matters

AMI says the capacity is meant to give cloud providers, enterprises and AI developers large-scale computing power closer to where they build and deploy AI. It describes the business as Compute-as-a-Service, meaning customers rent computing capacity instead of owning the hardware. AMI is also linked to Greenko's renewable generation and storage, and it says it aims to combine energy infrastructure with computing capacity.

For exam preparation, the story belongs under Science and Technology and Economy: AI infrastructure, data centres, energy demand and private investment in computing capacity.

Expansion plans and money figures

  • AMI plans to bring another 300 MW of capacity to market within the next 15 months.
  • The press release gives capital expenditure of more than $20 billion for this expansion, in addition to $6 billion already committed for the first 100 MW. GKToday reports the same figures without a currency symbol, so the dollar figures rest on the company's release.
  • The company describes a pipeline of about 400 MW of Compute-as-a-Service capacity across India, the United States, Europe and Malaysia. The three orders to date are the first committed tranche, and the rest is under development for phases in 2027 and early 2028.
  • Stated longer-term goals are 1 GW of Compute-as-a-Service capacity and a wider programme of 5 GW of powered AI data centres across India, the United States and Europe. These are targets, not completed capacity.

Exam-relevant terms

  • GPU (graphics processing unit): a processor widely used for parallel computing and AI workloads.
  • MW (megawatt): a unit of power equal to one million watts. Data-centre capacity is often measured in MW.
  • Liquid cooling: a method of removing heat from high-density computing equipment.
  • Rack-scale system: several computing components combined into one integrated deployment.
  • Compute-as-a-Service: computing capacity offered to customers as a service.

These definitions are background for understanding the news. No source says this topic will be asked in any exam.

What the evidence does not tell us

  • The split of the 20,000 GPUs and 70 MW between India and Malaysia, and the exact site locations, are not given.
  • The sources do not name customers or say how much capacity has been contracted to them.
  • Nearly all the figures come from AMI's own press release, and GKToday largely repeats them. The evidence includes no independent verification and no government statement.
  • The press release calls AMI one of the earliest and largest committed buyers of Rubin GPUs outside North America. That is the company's own claim and is not independently checked here.
  • The press release's wording on the first 100 MW capex is unclear, and the GKToday text omits the currency. Treat the money figures as company-stated.

FAQ

Is the 20,000 GPU order in addition to the earlier 9,000?

Yes. The 20,000 GPUs come from two further orders, added to the 9,000-GPU order announced on 25 August 2026. Together they take AMI's commitments to about 29,000 GPUs.

Have the GPUs been delivered?

No. Delivery of the new GPUs is scheduled for Q2 2027, which is after the 6 October 2026 as-of date.

← Back to Current Affairs