CoreWeave Enters India, Partners AdaniConneX For 240 MW Data Centre Capacity

What exactly is CoreWeave planning to build in Navi Mumbai, and who is it partnering with?

CoreWeave is planning a large AI-focused data centre campus in Navi Mumbai, Maharashtra. The project will provide 240 megawatts of data centre capacity. It matters because CoreWeave is entering India’s expanding AI infrastructure market and bringing specialised computing resources to customers there.

The campus will be built at AdaniConneX’s Taloja facility. It will contain three buildings, each designed for 80 MW. CoreWeave will be the sole tenant across all three buildings. The first phase is expected to become operational by mid-2028, while further capacity will be added in stages.

AdaniConneX is a 50:50 joint venture between Adani Enterprises and EdgeConneX, a US-based data centre operator. CoreWeave will deploy NVIDIA’s Vera Rubin platform and establish an India office. It can also choose to add another 240 MW, potentially doubling the campus to 480 MW.

What does “data centre capacity” measured in megawatts mean?

“Data centre capacity” measured in megawatts describes the electrical power a facility can deliver to its technology and supporting systems. A megawatt equals one million watts. The figure is therefore a useful measure of infrastructure scale, especially for facilities packed with power-hungry AI computers.

For example, CoreWeave’s planned 240 MW campus will supply power for GPU servers, storage, networking equipment, cooling systems, and other data centre operations. The number does not directly measure storage space, internet speed, or the number of AI models. It indicates the facility’s available power envelope.

A larger power envelope can support more computing equipment, although the exact workload depends on the hardware and design. Here, the campus is being designed for high-density AI infrastructure. Its direct liquid cooling and non-evaporative chilled water system are intended to help manage the heat created by that concentration of computing equipment.

How large is the planned facility, and how could its capacity grow from 240 MW to 480 MW?

The initial Navi Mumbai campus is planned to provide 240 MW of data centre capacity. It will consist of three separate buildings, each rated at 80 MW. The first phase is expected to be operational by mid-2028, and later capacity will come online in phases.

CoreWeave will be the sole tenant across the three buildings. Its agreement also includes the option to add another 240 MW at the same campus. That option would double the planned capacity from 240 MW to 480 MW. The article does not say exactly when the additional capacity would be built or how many extra buildings it would require.

This staged design lets the campus expand as demand develops. The initial 240 MW supports CoreWeave’s entry into India, while the possible 480 MW total leaves room for much larger AI workloads. CoreWeave plans to deploy NVIDIA’s Vera Rubin platform across the campus through its AI cloud platform.

What kinds of AI work will the NVIDIA Vera Rubin systems support, including training, inference, reasoning, and agentic workloads?

The NVIDIA Vera Rubin platform is planned for AI computing at CoreWeave’s Navi Mumbai campus. The article links it to four types of work: training, inference, reasoning, and agentic workloads. Together, these cover building AI systems, using them, enabling more complex problem-solving, and supporting systems that can perform multistep tasks.

Training uses computing power to develop or improve AI models from data. Inference is when a trained model produces an answer, prediction, or other output. Reasoning workloads involve handling more complex chains of thought or decisions. Agentic workloads refer to AI systems designed to carry out tasks through multiple steps, often using tools or services.

CoreWeave will provide these capabilities through its AI cloud platform. The planned infrastructure is intended for customers building and scaling AI in India. The article does not identify specific customers, models, or applications, but it says CoreWeave will work with customers and partners in India’s AI ecosystem.

Why does this project require high-density infrastructure, direct liquid cooling, and non-evaporative chilled water systems?

High-density AI infrastructure places many powerful GPUs and servers in a concentrated area. These systems require large amounts of electricity and turn much of that energy into heat. Without effective cooling, equipment could become less reliable or need to reduce performance. That is why the Taloja facilities are being designed specifically for dense AI workloads.

Direct liquid cooling brings coolant close to heat-producing components, allowing heat to be removed more efficiently than relying only on air. The campus will also use a non-evaporative chilled water system. It circulates chilled water to support cooling without consuming water through evaporation in the cooling process.

These systems match the project’s scale and intended workload. The article identifies direct liquid cooling for GPU data halls and non-evaporative chilled water as design features. Their inclusion shows that power delivery, heat removal, and water management are central engineering requirements for the planned AI campus.

What roles do CoreWeave, AdaniConneX, Adani Enterprises, and EdgeConneX play in the project?

CoreWeave is the AI cloud infrastructure provider leading the deployment and planned use of the campus. It will be the sole tenant across the three buildings, install NVIDIA’s Vera Rubin platform, and offer AI computing through its cloud platform. It also plans to open an India office and hire locally.

AdaniConneX is CoreWeave’s project partner at the Taloja facility. It is a 50:50 joint venture between Adani Enterprises and EdgeConneX. AdaniConneX is therefore the organisation connected to the facility and campus development, while its two parent companies provide the ownership structure behind that partnership.

The arrangement combines CoreWeave’s AI cloud expertise with AdaniConneX’s data centre platform. CoreWeave will initially use 240 MW and has an option for another 240 MW. The first phase is expected by mid-2028, with capacity added in phases as the campus develops.

Why are AI data centres becoming strategically important to India, and what basic resources—such as electricity, computing hardware, networks, and cooling—do they depend on?

AI data centres matter strategically because they provide the physical foundation for training and running AI systems. The article describes growing investment in India’s AI and data centre infrastructure. CoreWeave’s entry, Adani Group’s sovereign AI infrastructure plans, and partnerships involving Google, Meta, Uber, and Jabil show the sector’s expanding importance.

These facilities need several basic resources. Electricity powers GPUs, servers, storage, networking equipment, and cooling. Computing hardware performs AI calculations. Networks connect systems to users, partners, and other services. Cooling removes heat, especially in high-density GPU halls. The Taloja project addresses this need with direct liquid cooling and non-evaporative chilled water.

India’s planned infrastructure could support local AI customers and partners while strengthening domestic capacity. CoreWeave plans to hire locally and work with the country’s AI ecosystem. The article also points to future growth, including a possible expansion from 240 MW to 480 MW and wider investments by Adani Group and global technology companies.

Key Facts:

📌 CoreWeave plans 240 MW of capacity in Navi Mumbai.

📌 The campus will be built at AdaniConneX’s Taloja facility.

📌 CoreWeave can add another 240 MW later.

📌 A megawatt equals one million watts.

📌 Capacity measures a data centre’s available electrical power.

📌 The figure does not directly measure stored data.

📌 The campus will have three 80 MW buildings.