Vultr 20 MW AI Supercluster in Chennai: GPUs, Plans and Why It Matters, Explained
US cloud firm Vultr has launched a 20 MW AI supercluster in Chennai powered by AMD Instinct MI355X GPUs, with about 10,000 GPUs in the first configuration. India is already Vultr's largest market outside the US. Here is what was announced, what is still undisclosed, and what to watch next.
US cloud infrastructure provider Vultr has launched a 20 MW AI supercluster in Chennai, powered by AMD Instinct MI355X GPUs, The New Indian Express reported on 3 November 2026. The site is Vultr's largest AI deployment in India so far and is pitched at training large language models and foundation models. Here is what was announced, what was left undisclosed, and why the Chennai build matters for India's AI capacity race.
The 60-second version
- Launch: 20 MW AI supercluster in Chennai on AMD Instinct MI355X GPUs (The New Indian Express, 3 November 2026).
- Scale: About 10,000 GPUs in the initial configuration; among Vultr's two largest AMD AI deployments worldwide.
- Users: AI labs, startups, enterprises, research institutions and universities, for LLM and foundation-model training.
- India footprint: Vultr CMO Kevin Cochrane said India is already Vultr's largest market outside the US.
- Roadmap: Expand in 2027 with AMD MI455X and Helios rack-scale AI; target 600 MW of AI capacity in India by 2028 and gigawatt scale by 2029 (no firm timelines committed).
- Undisclosed: Investment amount; detail on existing Delhi-NCR, Mumbai and Bengaluru data centres.
What Vultr announced in Chennai
The New Indian Express (Express News Service, updated about 2:17 am on 3 November 2026) said Vultr launched the Chennai facility as its largest such deployment in India. The cluster is designed to support large language model and foundation model training.
Kevin Cochrane, Vultr's chief marketing officer, was quoted saying: "India is our largest market outside the US and it has highest footprint across the globe."
The paper also said Vultr's Chennai workforce could eventually reach 5,000 as the facility becomes operational and services expand.
GPUs and global ranking
The initial configuration is expected to house around 10,000 GPUs. That would place the Chennai site among Vultr's two largest AMD AI infrastructure deployments globally, according to the same report.
The GPUs named for day one are AMD Instinct MI355X. For 2027, Vultr plans to bring in AMD's MI455X GPUs and the Helios rack-scale AI architecture.
Growth targets for India
Beyond Chennai, Vultr is targeting 600 MW of AI infrastructure capacity in India by 2028 and gigawatt scale by 2029. The New Indian Express stressed that Vultr has not committed to firm timelines for those milestones, so treat them as company targets rather than booked projects.
Vultr already operates data centres in Delhi-NCR, Mumbai and Bengaluru, the paper said, but did not disclose investment figures for Chennai or operating detail for those sites.
Why this matters for India
India is racing to host more GPU capacity for local model training and for global cloud customers who want Asia-Pacific footprint. A 20 MW AMD-heavy site in Chennai adds coastal southern capacity alongside the established Mumbai and Bengaluru hubs, and signals that foreign hyperscale-style builders still see India as a core AI market outside the United States.
For startups and research labs priced out of scarce H100-class inventory, a large AMD MI355X pool is another procurement option, though pricing, access tiers and export-control limits were not covered in the launch report.
What to watch next
- Official Vultr and AMD fact sheets with rack counts, interconnect and customer onboarding dates.
- Investment and power purchase disclosures for the 20 MW Chennai site.
- 2027 MI455X / Helios expansion timelines if and when they firm up.
- Whether the 600 MW by 2028 target is backed by land, power and customer contracts.
- Local jobs: whether the stated path to about 5,000 Chennai staff materialises as the site ramps.
Chennai now hosts Vultr's biggest AI play in India. The missing pieces are the cheque size, the go-live schedule for the full 10,000-GPU pool, and whether the 2028-2029 megawatt targets harden into funded builds.
This explainer is based on The New Indian Express report dated 3 November 2026. Vultr had not published a detailed investment figure for the Chennai supercluster in that coverage. Figures and timelines will be updated if the company or AMD issue fuller disclosures.
Frequently asked questions
›What did Vultr launch in Chennai?
On 3 November 2026, The New Indian Express reported that US cloud infrastructure provider Vultr launched a 20 MW AI supercluster in Chennai powered by AMD Instinct MI355X GPUs. It is Vultr's largest such deployment in India so far.
›How many GPUs will the Vultr Chennai cluster have?
The facility is expected to house around 10,000 GPUs in its initial configuration, The New Indian Express reported, and will rank among Vultr's two largest AMD AI infrastructure deployments globally.
›Who will use Vultr's Chennai AI supercluster?
Vultr said the facility is designed for large language model and foundation model training, and will serve AI labs, startups, enterprises, research institutions and universities.
›What are Vultr's longer-term plans in India?
Vultr plans to expand capacity in 2027 with AMD's MI455X GPUs and Helios rack-scale AI architecture. It is targeting 600 MW of AI infrastructure capacity in India by 2028 and gigawatt scale by 2029, though The New Indian Express noted it has not committed to firm timelines.
›Did Vultr disclose the investment for the Chennai project?
No. The New Indian Express reported that Vultr did not disclose the investment involved in the Chennai project or details of its existing data centre operations in Delhi-NCR, Mumbai and Bengaluru.
›How big is India for Vultr?
Kevin Cochrane, Vultr's chief marketing officer, told The New Indian Express that India is Vultr's largest market outside the US and has the highest footprint across the globe among its markets.