The Story
Adani Enterprises and Jabil have signed a non-binding memorandum of understanding setting out the framework for a proposed 50:50 joint venture to manufacture and assemble artificial intelligence and data centre equipment in India. The venture remains subject to definitive agreements, approvals and customary conditions. It formalises the intent the two companies announced in June 2026 to build an AI and data centre infrastructure manufacturing ecosystem in India.
The proposed venture would manufacture and assemble AI racks, servers, storage and networking equipment, and make the equipment used to power and cool AI systems, including power distribution units, liquid-cooling systems and transformers. It would also provide installation and commissioning services. The companies say it is expected to have multi-gigawatt AI rack manufacturing capacity, serving hyperscalers, cloud providers, data centre operators and enterprises in India and abroad. No investment amount, location or timeline has been disclosed.
The companies project India's data centre capacity will reach 5 to 8 gigawatts by 2030, driven by AI adoption, cloud expansion and data localisation, with more than $50 billion in planned hyperscaler spending, and estimate the global market for AI computing infrastructure at more than $3 trillion over the next seven years.
"At Adani, our vision is to expand up the AI value chain, evolving from powering and hosting AI infrastructure to engineering and manufacturing the critical systems that will define the next generation of computing," said Gautam Adani, chairman of the Adani Group.
Jabil chief executive Mike Dastoor said the company believes India will become increasingly important for both domestic and global AI infrastructure demand, and that the venture would let Jabil bring its rack-level data centre products to more hyperscalers in India and around the world.
Jabil, a US engineering and manufacturing company listed on the New York Stock Exchange and a supplier to Apple, expects fiscal 2027 revenue of $44.5 billion. The Adani Group builds data centres through AdaniConneX, its joint venture with EdgeConneX, and has pledged to invest $100 billion in AI data centres over the decade to 2035.
Why It Matters
For two years the global conversation about AI infrastructure has been about chips. The constraint now runs further along the chain. A working AI data centre needs racks that integrate dozens of accelerators with networking, power and liquid cooling, and electrical equipment to feed them at a scale data centres were never designed for. Assembling and testing those racks is specialised work, done today mostly by a handful of contract manufacturers in Taiwan, China, Mexico and the United States, and the power equipment behind them, transformers above all, has become one of the scarcest parts of the build-out, with lead times stretching into years.
That is the part of the chain this venture targets. Jabil brings the engineering and manufacturing experience: it already builds rack-level data centre systems and makes components for Apple. Adani brings what a manufacturer in India finds hardest to assemble alone: land, power, approvals and, crucially, demand. Gautam Adani's own description of the plan is a move up the value chain, from generating the power and building the data centres to making the equipment that goes inside them.
India wants exactly that. Servers fall within the government's production-linked incentive scheme for IT hardware, and data localisation rules and the country's sovereign AI ambitions mean more computing will have to sit inside its borders. A domestic supplier of racks and power systems shortens supply chains for that build-out, and gives India something to export into a global market the companies size in the trillions of dollars.
The Strategic Read
The first thing to hold on to is that nothing has been committed. A non-binding MoU sets out a framework; the investment, the factory location, the capacity in units and the timeline are still to come, and the $3 trillion and 5 to 8 gigawatt figures describe markets, not this venture. The step from June's statement of intent to a signed framework in October is real progress. The definitive agreement is the one that will matter.
The second is what made in India will mean here. The venture would assemble racks and servers and make power and cooling equipment. The GPUs, processors and high-bandwidth memory at the heart of those racks are made by a handful of companies, largely in Taiwan and South Korea, and nothing in this deal changes that. India's share of the value would be integration, testing and the electrical and thermal systems around the chips. That is still meaningful, and the power and cooling side may be the more strategic half, given how scarce transformers and liquid-cooling capacity have become. But it is a different thing from making AI hardware in the sense the chip industry would use the phrase.
The third is the customer. The most obvious early buyer for an Adani-Jabil rack is an Adani data centre. AdaniConneX is building AI campuses in Visakhapatnam and Noida, including the hub Google plans in Visakhapatnam, and the group has pledged $100 billion to AI data centres through 2035. That gives the venture volume from day one, which is how most manufacturing ventures of this kind survive their early years. It also raises a question the companies have not addressed: whether rival data centre operators and hyperscalers will buy critical equipment from a venture half-owned by a competitor in their own market.
For Jabil, the logic is diversification. US manufacturers have spent several years shifting capacity away from China, and an Indian partner with power, land and a captive customer lowers the risk of building here. For Adani, it is a familiar strategy: own the whole chain, from the power plant to the server rack, and capture margin at each step. Whether that integrated model works as well for precision electronics as it has for ports, power and airports will depend far more on Jabil's half of the partnership than on Adani's.
For daily, sharp analysis of the biggest moves in the Indian business and startup ecosystem, follow StartupFox.
