India’s third-largest private-sector lender is doubling its data-centre exposure. The harder question is how a bank underwrites an asset it cannot easily seize and sell.
This is a Bloomberg interview read through an Economic Times mirror; Bloomberg’s own page is paywalled to this publication. The executive is named and quoted directly.The doubled rupee figure below is this publication’s arithmetic on today’s base. Axis has stated no absolute target, and nothing here assesses the bank’s credit quality. Nothing here is investment advice.
Axis Bank, India’s third-largest private-sector lender by assets, plans to double its data-centre lending over the next three years. Vijay Mulbagal, Axis’s group head of wholesale banking coverage, told Bloomberg — in an interview read here via an Economic Times mirror, because Bloomberg’s own page is paywalled to this publication — that data-related investments in India are growing, and companies are resorting to borrowings alongside equity to fund larger capacities.
The current base is modest. Data centres account for a little over two per cent of Axis’s corporate loan book. Economic Times puts that at more than Rs 86.88 billion — approximately $907 million, ET’s own currency conversion derived from the bank’s first-quarter figures. Axis serves about a dozen data-centre clients today. The bank has also approached Singapore-based data-centre companies about financing, though whether those approaches have produced transactions is not established.
If the corporate book held flat and the segment share simply doubled, this publication’s arithmetic — not any stated Axis target — would put the exposure at roughly Rs 17,376 crore. Axis has given no absolute rupee figure. The corporate book is not static over three years either, so the figure is an order of magnitude rather than a commitment.
The key insight: Ashwini Kumar Tewari, State Bank of India’s managing director, said last month that cash-flow-based lending would become increasingly important for industries like data centres, because such businesses may not have the traditional collateral that bank finance requires. That is the structural problem in one sentence.
The two objects are not the same thing. One is a single bank’s exposure; the other is an estimate of what the whole sector needs to borrow.
The Structural Read
Traditional corporate lending is built on a simple assumption: if the borrower fails, the lender takes the asset and sells it. That is what collateral means in practice. The asset’s forced-sale value sets the ceiling on what a bank lends.
A data centre breaks that assumption. The building, the cooling systems, the racks — these have some residual value. But the economic value of the facility is not the shell. It is the capacity contracts, the grid connection, and the land rights. Strip the tenants out and the resale value of what remains is not the loan.
So the underwriting question changes shape entirely. A lender stops asking what this is worth in a forced sale. It starts asking who is paying for the capacity, for how long, under what terms, and what happens if they stop. That is cash-flow-based lending — and it is a different discipline from the asset-backed lending that still dominates most Indian corporate bank books.
Mulbagal’s own phrasing is telling: borrowings come “apart from bringing in equity.” The debt is the supplement. The equity balance sheet is the base. This matters because the announced ambitions in Indian data-centre infrastructure are enormous. Gautam Adani has outlined plans to invest $100 billion in AI-ready data centres by 2035. Mukesh Ambani’s Reliance has said it will invest as much as $110 billion in AI-related infrastructure over seven years.
These are announced ambitions cited as context — not funded commitments, and this publication is not implying that either borrows from Axis.
Even so, the scale comparison is instructive. NaBFID estimates the Indian data-centre sector will need about Rs 1 lakh crore in funding through March 2031 and has already sanctioned more than Rs 3,000 crore each to at least four data centres. Even at a doubled base, one of India’s largest private lenders would hold around a sixth of that single estimate. The two are different objects — one bank’s exposure against a sector-wide funding need — but the order of magnitude is the point. Bank debt is not the primary engine here.
SBI Managing Director, reported speech via Economic Times
Reported speech, not a direct quotation: Economic Times records that Tewari said last month that cash-flow-based lending would become increasingly important for new-age industries such as data centres, as such businesses may not have the traditional collateral required for bank finance.
— Ashwini Kumar Tewari, State Bank of India managing director
The Global Backdrop
The United Nations Conference on Trade and Development reported that more than a fifth of the value of greenfield projects globally in 2025 was linked to data centres, with announced investments exceeding $270 billion. KPMG puts India’s data-centre value-chain opportunity at an estimated $90 billion by fiscal 2035.
The constraints are physical, not financial. Power, water, and large parcels of land each carry delay and cost risk. Residents in Thane, near Mumbai, have already protested a proposed Amazon data centre. A proposed 20-year tax holiday for foreign companies providing global data-centre services from India has not been enacted.
Three Implications
IMPLICATION 1 — THE UNDERWRITING CAPABILITY GAP Doubling exposure is a capital allocation decision. Building the internal capability to assess tenant concentration risk, contract duration, and power offtake agreements is a different kind of work. That capability is either built in-house or bought in, and neither route is free. This publication is describing the work the shift implies, not predicting which lenders do it.
IMPLICATION 2 — BANK DEBT AS INFRASTRUCTURE SIGNAL A major private lender increasing a sector allocation is a datapoint other lenders can read, though nothing in the report establishes that it has moved anyone else’s view. NaBFID’s debt-fund positions already signal institutional appetite. Axis’s stated direction adds a second data point. What that does to the cost of capital for contracted data-centre capacity in India is not something this report establishes, and this publication is not forecasting it.
IMPLICATION 3 — THE PHYSICAL CONSTRAINT IS THE BINDING VARIABLE On the constraints named in the report, the binding ones are physical rather than financial: power connections, land with grid access and permitting timelines. A data centre that stalls on a grid connection is not a performing loan, regardless of how well the contracts were written. Lending against that profile puts physical infrastructure risk on a balance sheet that has not historically carried much of it.
Business Engineer Framework
Harness Theory
Axis is not building AI infrastructure. It is harnessing the economics of those who are. Harness Theory maps how companies capture value from AI-era investment without owning the underlying capability. The Map of AI shows which layer each player sits in — and which layers the money actually flows through.
Axis Bank doubling its data-centre book is a footnote against the scale of India’s AI infrastructure ambitions. What it is not a footnote to is the structural shift it requires inside the bank. Moving from collateral-based to cash-flow-based underwriting is a genuine capability change, and it is the part of this story that does not scale with the size of the loan book.
Sources: Economic Times (Bloomberg interview mirror), October 2026. Original interview by Bloomberg, paywalled to this publication. Figures attributed to Economic Times where ET’s own conversion or derivation; arithmetic on doubled base is this publication’s own calculation from ET’s Q1-derived figures, not a stated Axis target. NaBFID, UNCTAD, and KPMG figures cited as reported. Adani and Reliance investment figures are announced ambitions, not funded commitments. Nothing in this article constitutes credit assessment or investment advice.
Every detail above comes from a Bloomberg interview with Axis Bank, read here through an Economic Times mirror of 1 October 2026. Bloomberg’s own page is paywalled to this publication and has not been read directly. The Axis executive, Vijay Mulbagal, and the State Bank of India managing director, Ashwini Kumar Tewari, are named and quoted in that text.Two figures are derived rather than reported.
The $907 million is Economic Times’ conversion of Rs 86.88 billion, which Economic Times itself calculated from the bank’s first-quarter corporate book share. The Rs 17,376 crore doubling figure is this publication’s arithmetic on that base. Axis has stated no absolute target, and its corporate book will itself change over three years, so the doubled figure is an order of magnitude rather than a projection.
The comparison with NaBFID’s estimate sets two different objects side by side. One is a single bank’s exposure; the other is an estimate of what the whole Indian data-centre sector needs to borrow through March 2031. They are not the same measure and the comparison is offered only for scale.The Adani and Reliance figures are announced investment ambitions cited as context in the same report.
Nothing above claims they are funded, committed, underway, or financed by Axis Bank, and neither company is described as an Axis client. The 20-year tax holiday is a government proposal, not enacted law.Nothing above assesses Axis Bank’s credit quality, loan book or risk management, and no view is offered on whether this lending is prudent. Also absent: which data-centre clients, loan pricing or tenor, any stress or default figure, and whether the Singapore approach produced anything. Axis Bank is a listed company. Nothing above predicts anything, and nothing here is investment advice.
Gennaro is the creator of FourWeekMBA, which reached about four million business people, comprising C-level executives, investors, analysts, product managers, and aspiring digital entrepreneurs in 2022 alone | He is also Director of Sales for a high-tech scaleup in the AI Industry | In 2012, Gennaro earned an International MBA with emphasis on Corporate Finance and Business Strategy.
Scroll to Top
Discover more from FourWeekMBA
Subscribe now to keep reading and get access to the full archive.
