Adani Enterprises and US-based Jabil have signed a non-binding memorandum of understanding to establish a 50:50 joint venture for manufacturing artificial intelligence and data-centre equipment in India. The proposed venture would make AI racks, servers, storage and networking equipment, as well as power distribution and liquid-cooling systems, placing industrial hardware at the centre of India’s attempt to expand its digital infrastructure.
The agreement is significant because data centres are usually discussed as buildings filled with servers, while the proposed partnership points to a wider system: energy supply, advanced computing, cooling, manufacturing and network connectivity operating together. The companies say the venture will develop multi-gigawatt manufacturing capacity for domestic demand and exports. However, the proposal remains subject to definitive contracts, regulatory approvals and other customary conditions. No final investment, location, production schedule or employment number has been provided in the supplied announcement.
That distinction matters. The MoU establishes an intended direction rather than a completed manufacturing project. Its value for India will depend on whether the companies move from a broad industrial proposition to specific plants, technology lines, procurement commitments and operating capacity. For now, the announcement confirms that the partners want to combine Jabil’s engineering and manufacturing capabilities with Adani’s infrastructure, energy and data-centre presence.
The proposed joint venture follows an announcement in June 2026 about the companies’ plans to develop an AI and data-centre infrastructure manufacturing ecosystem in India. The new agreement adds a manufacturing layer to that earlier plan. Instead of focusing only on operating data centres or supplying power to them, the partnership is intended to produce the equipment that allows those facilities to function.
That equipment is becoming more demanding as AI workloads expand. AI racks and servers require high-density computing, while liquid-cooling systems are designed to manage the heat generated by concentrated computing hardware. Power distribution equipment is equally important because data-centre reliability depends not only on computing capacity but also on the ability to deliver stable electricity to large installations.
The companies have not disclosed the proposed venture’s manufacturing locations or the specific technologies that would be produced in India. They have said it would serve hyperscalers, data-centre operators, cloud service providers and enterprises. This customer base spans the largest global computing companies as well as businesses that increasingly depend on cloud systems and data-intensive services.
The industrial proposition is therefore closely connected to the physical growth of India’s data-centre market. According to estimates cited by the companies, India’s data-centre capacity could reach 5-8 gigawatts by 2030, driven by increased AI adoption, cloud expansion and data localisation. The range itself indicates the scale of the opportunity, but it is also a projection rather than an operating capacity figure. The supplied announcement does not provide a current national capacity baseline or identify how much of the projected requirement would be met by the proposed venture.
The companies also estimate the global market opportunity for AI computing infrastructure at more than $3 trillion over the next seven years. That figure is presented as an industry opportunity estimate, not as the value of the joint venture or a guaranteed revenue pool. It nevertheless explains why equipment manufacturing is becoming a strategic part of the data-centre business. As computing demand grows, the ability to assemble or manufacture racks, servers, cooling systems and power equipment can influence project timelines, supply-chain resilience and the cost of expanding capacity.
For India, domestic manufacturing could potentially address a structural issue in digital infrastructure: the need to coordinate physical facilities with the equipment installed inside them. Data-centre construction is not complete when a building is ready. The facility also requires computing hardware, electrical systems, cooling infrastructure and network equipment that must be procured, delivered, integrated and maintained. A domestic production base would place more of those activities within the country, although the announcement does not establish the extent of local sourcing or the components that would continue to be imported.
The proposed venture also brings energy into the centre of the discussion. Adani Group chairman Gautam Adani said the AI revolution would require “an entirely new industrial architecture where energy, hyperscale infrastructure, advanced computing and precision manufacturing converge”. The statement reflects the operational relationship between data centres and power systems: computing capacity needs dependable electricity, and high-density computing increases the importance of cooling and power management.
The Adani Group has said it has committed $100 billion to develop 5GW of green-energy-powered, AI-ready data centres by 2035. The supplied report does not provide a project-by-project breakdown of that commitment or clarify how the proposed manufacturing venture would be integrated with those facilities. It does, however, show how the group is positioning manufacturing alongside energy, infrastructure and data centres rather than treating each as a separate business.
Jabil’s role is based on its existing engineering and manufacturing capabilities. Chief executive Mike Dastoor said the partnership would combine Jabil’s manufacturing expertise with Adani’s infrastructure and energy capabilities to expand rack-level data-centre solutions in India and overseas. Jabil reported revenue of $36 billion in fiscal 2026 and has expanded its data-centre capabilities through acquisitions involving critical power management and liquid-cooling technologies, according to the report.
This combination is important at the rack level. A data centre’s performance depends on the interaction of servers, storage, networking, power distribution and cooling. Producing or integrating those systems under a coordinated industrial arrangement could reduce the distance between equipment design and data-centre deployment. Whether it will do so in practice depends on the final structure of the venture, the approvals it receives and the scale at which facilities are built.
The policy environment is also becoming more supportive of data-centre investment. The Union Budget 2026 announced a tax holiday for data centres until 2047, while the Digital Personal Data Protection Act, 2023, provides a framework for personal data protection. The tax measure is intended to improve the long-term investment environment for data infrastructure, while the data-protection framework is part of the legal context in which companies store and process information.
These measures address different parts of the data-centre ecosystem. Fiscal incentives can influence the economics of capital-intensive projects, but they do not by themselves resolve questions about land, electricity, water, transmission capacity, fibre connectivity, environmental approvals or local infrastructure. The supplied announcement does not specify how the proposed joint venture will address those requirements. Its immediate focus is manufacturing equipment rather than the full regulatory or resource profile of individual data-centre sites.
The partnership also highlights the relationship between data localisation and physical infrastructure. When data is required or encouraged to remain within national boundaries, demand for domestic data-centre capacity can increase. That demand creates requirements not only for buildings and servers but also for power systems, cooling technology and maintenance capabilities. The companies explicitly link India’s projected capacity growth to data localisation, along with AI adoption and cloud expansion.
The scale of the proposed manufacturing plan is described as multi-gigawatt capacity, but the term is not accompanied by a plant count, annual output figure or commissioning timetable. That leaves important questions unanswered. It is not yet clear where production would take place, what proportion of equipment would be intended for India, how exports would be organised, or how the venture would measure progress from agreement to operational capacity.
The same uncertainty applies to the relationship between manufacturing and urban development. Data centres are major infrastructure users, even when they are not visible to most residents. They require industrial land, high-capacity electricity connections, cooling systems, access roads, telecommunications networks and, depending on the design, substantial water-management arrangements. A hardware manufacturing ecosystem could support this expansion, but it could also increase the need for coordinated planning across industrial, energy and digital infrastructure agencies.
The announcement therefore represents more than a corporate tie-up, but less than a confirmed manufacturing build-out. It establishes a proposed institutional arrangement between a major Indian infrastructure group and a global electronics manufacturer. It also shows that India’s AI ambitions are moving into the physical systems required to run computing at scale.
The next test will be whether the non-binding MoU becomes a definitive agreement and then an operating industrial project. The milestones to watch are regulatory approvals, final contracts, locations, investment commitments, manufacturing timelines and clarity on the equipment to be produced. Until those details emerge, the agreement is best understood as a strategic signal: India’s data-centre expansion is creating demand for an integrated ecosystem of energy, buildings, computing hardware, cooling and power management, and companies are beginning to organise around that requirement.
