AI, chips, defence to drive Rs 16 lakh cr in India’s next capex cycle: Union Bank report
AI, semiconductors and defence are expected to attract investments of around Rs 5.9 lakh crore, Rs 5.27 lakh crore and Rs 5.07 lakh crore, respectively, from public and private sources.
India’s next wave of corporate investment is likely to be led by artificial intelligence (AI), semiconductors and defence, with indicative public and private capex of about Rs 5.9 lakh crore, Rs 5.27 lakh crore and Rs 5.07 lakh crore respectively—close to Rs 16.24 lakh crore across the three sectors. A Union Bank of India research report says this spending will steadily shift banks’ credit opportunities towards battery storage, semiconductor packaging, green hydrogen projects with assured buyers and defence component suppliers over the next three to five years.
The report suggests that while traditional sectors such as steel, cement, automobiles and conventional manufacturing will remain the base of the capex cycle, newer industries are becoming large enough to create meaningful lending demand. It identifies renewables, energy storage, data centres, semiconductors, electronics, defence and advanced manufacturing as the key areas where banks can find future credit opportunities.
AI is projected to attract around Rs 5.9 lakh crore, semiconductors about Rs 5.27 lakh crore and defence roughly Rs 5.07 lakh crore in indicative public and private investments. Within these sectors, the financing opportunity is not limited to large plants or flagship projects. The report expects demand for loans to emerge in semiconductor advanced packaging and assembly, battery recycling, lithium iron phosphate (LFP) and sodium-ion battery manufacturing, storage infrastructure, and suppliers making components for defence companies.
Green hydrogen is another area where banks may see new business, but the report links it to a crucial condition: projects must have assured offtake agreements. In other words, lenders are likely to favour hydrogen facilities that already have buyers committed to purchasing the output, rather than projects relying only on uncertain future demand. Data Centers move closer to users The report also expects data centre demand to evolve as AI adoption increases.
Instead of concentrating only in large hyperscale facilities, growth is likely to extend towards edge data centres and colocation facilities as AI inference—the process of running trained AI models for users—expands. This shift matters for lenders because data centres require heavy upfront investment in land, power, cooling systems, servers and network infrastructure. As AI workloads move closer to businesses and consumers, smaller but more distributed facilities could create a new layer of financing demand.
4 billion aerospace metals market Lenders face long payback risk The report is clear that these opportunities come with risks. Technology can become outdated quickly, revenues may remain uncertain, payments can be delayed, and many projects take years before they generate adequate returns. Semiconductor projects, for instance, typically have payback periods of five to eight years, while space projects may take seven to ten years.
These timelines can exceed conventional loan tenures, making standard project financing difficult. AI infrastructure carries an additional complication: rapid technology cycles. The report notes that graphics processing unit, or GPU, clusters may require upgrades every 18 to 24 months, which can affect both the useful life of assets and the economics of loans backed by them. Debt is available, but not everywhere Despite these concerns, the report finds that ten of the twelve emerging sectors it examined already have identifiable debt-financing routes.
The exceptions are AI and space, which remain largely dependent on equity funding because of high technology risk and long payback periods. The report differentiates sectors by the kind of finance they are likely to need. Renewable energy, semiconductors, green hydrogen and critical minerals are seen as suitable for project finance and funding led by development finance institutions. Electric vehicles and drones, by contrast, could benefit more from asset financing and lending by non-banking financial companies.
Banks need specialised products For banks, the report recommends building specialised lending capabilities in semiconductors, hydrogen, data centres and battery storage. It also suggests loan structures that match the cash-flow realities of these businesses, including milestone-linked disbursements, receivables financing, equipment leasing and green loans. Such products could reduce lender risk by linking fund release to project completion or operational milestones, while receivables financing can help suppliers manage delayed payments from larger buyers.
Equipment leasing may also suit sectors where machinery becomes outdated quickly. India’s renewable energy capacity stood at 288.6 GW, against a target of 500 GW by 2030. The scale of the storage opportunity is also growing. The government has projected a requirement of 208 GWh of battery energy storage systems by 2030 to help integrate a rising share of renewable power into the grid. Meanwhile, the report projects that India’s electric vehicle and battery storage market could expand from USD 34 billion to USD 113 billion by 2030.
That growth is expected to create financing demand not only for manufacturers, but also for charging infrastructure, battery plants, recycling units and supply-chain companies. For banks and investors, India’s lending opportunities are expanding beyond traditional infrastructure and manufacturing. As new sectors grow, funding is likely to go to projects and suppliers with steady income, reliable customers and manageable technology risks.
For companies in storage, chips, hydrogen, data centres and defence manufacturing, access to capital may increasingly depend on showing lenders a clear path to cash flows—through assured offtake agreements, milestone-based execution and realistic repayment timelines.