Adani and Jabil Want to Build AI Hardware in India. Who Will Capture the Value?

 

TECHNOLOGY • MANUFACTURING • DATA CENTRES

A proposed 50:50 joint venture would move Adani and Jabil further into the equipment layer of the AI economy. The commercial test is whether India can manufacture critical systems—not merely host data centres.

By Murthy ASR  |  14 October 2026  |  Biz Masala Signature Story

EDITORIAL VISUAL • Illustrative artwork created for this story; not a photograph of a specific facility or transaction.

THE CENTRAL QUESTION  A proposed 50:50 joint venture would move Adani and Jabil further into the equipment layer of the AI economy. The commercial test is whether India can manufacture critical systems—not merely host data centres.

A non-binding agreement, not a finished factory

Adani Enterprises and US-based Jabil announced on 13 October 2026 that they had signed a non-binding memorandum of understanding outlining a proposed 50:50 joint venture to manufacture and assemble AI and data-centre equipment in India. The proposed product scope includes AI racks, servers, networking, power and cooling systems.

The companies said the proposed venture would serve demand from hyperscalers, neocloud providers, data-centre operators and enterprises. The MoU remains subject to definitive agreements, applicable approvals and customary conditions. Until those steps are completed, the proposal should not be described as an operational joint venture or a committed factory with a confirmed capex figure.

Why the hardware layer matters

AI is often discussed as software, models and applications. In practice, the expansion of AI services depends on a physical stack: land, electricity, cooling, server racks, networking equipment, semiconductors and specialised manufacturing. Delays or cost overruns in any part of that stack can constrain the entire project.

Adani brings Indian infrastructure experience and a stated ambition to move up the AI value chain. Jabil brings global engineering, supply-chain and manufacturing capabilities. The partnership’s potential lies in combining local infrastructure and market access with manufacturing expertise. The risk is that final assembly alone may capture less value than deeper component engineering, design, testing and localisation.

The numbers are forecasts, not orders

The companies’ announcement referenced industry forecasts of 5–8 gigawatts of Indian data-centre capacity by 2030 and more than US$50 billion in planned hyperscaler spending across data-centre, cloud and AI ecosystems. These figures describe an anticipated market opportunity; they are not orders booked by the proposed venture and should not be treated as guaranteed revenue.

The announcement also cited a global AI compute infrastructure opportunity exceeding US$3 trillion over seven years. Such estimates are useful for understanding strategic intent, but the scope and assumptions behind market forecasts matter. The addressable market is not the same as the revenue available to one manufacturer.

Five questions that will decide whether the deal matters

Where will the facility or facilities be located, and when could production begin? What capital will each partner commit? Which components will be sourced locally, and which will be imported? Who owns product design, intellectual property and customer relationships? Finally, will the venture export equipment, or primarily supply Indian projects?

These are not minor details. They determine whether the venture creates a durable domestic manufacturing capability or becomes a final-assembly operation dependent on imported high-value components. They also shape the quality of jobs created, supplier opportunities and the resilience of India’s AI infrastructure.

The Biz Masala View

The Adani–Jabil proposal is strategically interesting because it targets the equipment layer of AI, where infrastructure spending can translate into manufacturing demand. But a non-binding MoU is the start of a commercial process, not proof of industrial transformation. The accountability test is straightforward: definitive agreements, disclosed capex, a location, a commissioning timetable, credible customer commitments and evidence of domestic value addition.

The proposed venture: status and evidence checklist

Issue	Current position	Evidence to watch
Structure	Proposed 50:50 JV	Definitive agreement and incorporated entity
Status	Non-binding MoU announced 13 Oct 2026	Approvals and conditions satisfied
Products	Racks, servers, networking, power, cooling	Product list, specifications and production capacity
Market	Companies cite fast-growing AI/data-centre demand	Customer contracts and order backlog
Local value	Manufacturing and assembly proposed in India	Local sourcing, engineering, exports and jobs

 

The numbers in context

Gigawatts, forecast range cited in the companies’ 13 Oct 2026 announcement; not JV capacity or confirmed orders.

SECOND EDITORIAL VISUAL • Conceptual illustration of the business mechanism discussed above.

Source notes and verification

The article distinguishes company announcements and reported facts from analysis. Figures described as forecasts, proposals or management assessments are not presented as realised outcomes.

·       Jabil official announcement, 13 October 2026

·       Reuters, 13 October 2026

Editorial note: Verify the latest exchange filings and original documents immediately before publication, particularly where a proposal remains non-binding or a legal process is unresolved.

 

BUSINESS  •  MARKETS  •  PEOPLE  •  INSIGHT

The Biz Masala View: follow the evidence, not just the announcement.