SBI Cards: India's Credit Card Bet, and Whether the Bet Still Works

I. Introduction & Roadmap

On the morning of March 16, 2020, a bell rang on Dalal Street for a company that had spent twenty-two years as somebody else's joint venture. India's markets were in freefall — the World Health Organization had declared a pandemic five days earlier, and global equities were pricing in the end of ordinary commerce. SBI Cards and Payment Services Limited, which had priced its initial public offering at ₹755 a share after a book that was covered more than twenty-six times, opened at ₹658.¹ It was, in the language of investment banking, a broken IPO on day one. It was also, in a more literal sense, a company whose entire business model — persuading Indians to spend money in shops, restaurants, airports and hotels — had just been made illegal by a national lockdown that would begin eight days later.

More than six years on, an investor who bought at that offer price is still underwater. That single fact is the reason this story is worth telling carefully rather than enthusiastically.

SBI Card is the only pure-play, standalone credit card issuer of consequence in India. It is not a bank. It is a non-banking financial company that happens to carry the most powerful retail brand in Indian finance on the front of its plastic. Its promoter, State Bank of India, holds 68.58% of the equity and serves more than 500 million customers through in excess of 23,000 branches — a customer list no competitor can buy and no fintech can replicate.²³ That list is the bull case in a single sentence.

The bear case is also a single sentence, and it sits directly underneath: SBI Card funds itself in the wholesale market, at rates set by bond investors and rating committees, while every one of its major competitors — HDFC Bank, ICICI Bank, Axis Bank, Kotak Mahindra — funds the identical business with current-account and savings-account deposits gathered at a fraction of the cost.

The company owns the best distribution in the industry and the worst balance sheet in the industry, and the entire investment question is whether the first outweighs the second across a full cycle. Everything else — the rewards programmes, the co-brand partnerships, the digital onboarding funnel, the market-share tables that analysts update monthly — is detail layered on top of that one structural asymmetry.

It is also worth being clear at the outset about what kind of company this is not. SBI Card is not a fintech, though it is often bundled with them. It does not run a payments network; it rents one. It does not hold deposits, sell insurance at scale, or cross-subsidise from a corporate lending book. It does one thing: it lends unsecured money to Indian consumers at high rates and collects a toll on their spending. When that single product line performs, the return on equity is excellent. When it does not, there is nothing else in the building.

Between late 2023 and early 2026, that single-product concentration stopped being theoretical. India's unsecured lending boom turned, the Reserve Bank of India raised the capital cost of exactly the loans SBI Card makes, delinquencies climbed for six straight quarters, and profit fell 20% in a single financial year.⁴ Management said, repeatedly and sincerely, that the worst was close. It was not, for a considerable while. Then, in FY26, it genuinely was: gross NPAs fell to 2.41% from 3.08%, profit rose 13% to ₹2,167 crore, and by the June 2026 quarter gross NPAs stood at 2.04% and gross credit cost at 6.5% — the cleanest asset-quality print the company had shown in more than three years.⁵⁶

So the company has come out the other side.

The stock has not: at ₹646 on August 30, 2026, it trades roughly 14% below where it was priced in March 2020.⁷ Something in the middle of that sentence deserves explanation, and this piece is an attempt to supply it.

The route runs as follows.

First, the origins — a GE Capital joint venture, a private-equity handoff, and the worst-timed listing in modern Indian capital markets. Then the economics of a credit card company, explained plainly, because almost every argument about SBI Card collapses into an argument about how it funds itself. Then the competitive map, including the regulatory rewrites of 2022 to 2026 that reshaped the game. Then the distribution question — moat or rented advantage. Then management, ownership and capital allocation, where the record is thinner than the reputation. Then the reckoning itself, tested against what management actually said on each quarterly call. Then RuPay on UPI, which is genuinely undecided. And finally the bull and bear cases, the risks that matter, and the two or three numbers worth watching from here.

↑ Back to Top

II. Origins: A GE Capital JV, a Private-Equity Handoff, and a COVID-Week IPO (1998–2020)

In 1998, the Indian credit card was a curiosity — an object carried by senior executives and NRIs, accepted at a few hundred thousand merchant terminals, and viewed by most banks as a cost centre attached to the current-account relationship. State Bank of India, the nation's balance sheet, had the customers but not the machinery: card issuance is a business of scorecards, collections infrastructure, fraud models and merchant settlement, none of which sat naturally inside a public-sector bank's branch hierarchy.

So SBI did what a number of Indian institutions did in that era. It found a foreign partner who had already industrialised the thing. GE Capital, then the most respected consumer-finance operation on earth, brought the underwriting technology and the operating discipline; SBI brought the name and the customers. The joint venture that resulted was India's first dedicated credit card company — an entity built on the premise that specialisation beats a bank's card desk. When GE finally sold out, its chairman noted that State Bank had been a partner "over the last 19 years," dating the relationship to 1998.⁸

What GE brought was not capital — SBI had plenty of that — but a manufacturing mindset applied to consumer credit. GE Capital's consumer finance arm ran card portfolios across a dozen countries on a common template: score the applicant, price the risk, set the limit algorithmically, monitor the vintage, and treat collections as a production line rather than a legal process. Those disciplines, imported into a joint venture in Gurugram, are the reason SBI Card entered the 2010s with underwriting infrastructure that most Indian banks' card desks did not have. It is a genuine inheritance, and it is worth remembering when assessing the company's later credit performance: the operating machinery was not built in a hurry.

The GE era ended for reasons that had almost nothing to do with India. After the financial crisis, General Electric spent the better part of a decade dismantling GE Capital, and by 2017 the Indian card joint venture was simply an asset on a disposal list. On July 21, 2017, State Bank of India and The Carlyle Group announced definitive agreements to jointly acquire GE Capital's entire stake in the two entities that operated SBI Card, leaving SBI with 74% and Carlyle with 26%; Carlyle's own portfolio records date the completed investment to December 2017.⁸⁹

This is the moment that set up everything afterwards, and it is worth pausing on. SBI could have bought GE out alone. It chose instead to bring in a global private-equity sponsor — which tells you the bank knew it needed something other than capital. Carlyle's job was to professionalise the company for a public listing: sharpen the reporting, build the equity story, and put an independent commercial mind in the room during a growth phase. For roughly two and a half years, that is what happened. The card base grew, the digital sourcing engine was built out, and the company was readied for the market.

Then came the listing. The IPO was structured as a very large offer for sale, raising approximately ₹10,340 crore — one of the largest in Indian history at the time — at a price band topping out at ₹755 per share.¹ Institutional demand was extraordinary. The book was covered more than twenty-six times.

The pitch was clean and, on the numbers then available, entirely defensible: India had roughly one credit card for every twenty-five people; the United States had roughly four in five adults carrying one; SBI Card sat on the largest customer funnel in the country; the arithmetic did the rest. Retail investors queued for allotments. Employees and SBI shareholders were given reserved portions. In the last week of February 2020, it was the most wanted piece of paper in Indian finance.

The offer closed in the first week of March 2020. Between the closing of the book and the listing date, the pandemic arrived. The stock listed on March 16, 2020 at ₹658, a discount of roughly 13% to the offer price, and kept falling — within weeks it was down by roughly a third from where it had priced.¹

Then the real problem started. India's national lockdown began on March 24, 2020, and for the following two months the transaction categories that generate the most valuable card spending — travel, hotels, restaurants, entertainment, department stores, fuel — were simply switched off. A credit card issuer in a lockdown faces a peculiar double bind: revenue collapses because nobody is spending, while credit risk rises because the same customers have lost income and cannot repay what they already borrowed. Regulatory moratoria on repayment, introduced to protect borrowers, made the second problem harder to measure in real time. Gross NPAs, which had been 1.35% in the June 2020 quarter, climbed toward 4.9% by the March 2021 quarter, and management told investors during the second wave that it was simply too early to take a call on where losses would settle.³⁹

That episode matters for a reason beyond narrative. It established the company's disaster base rate. Whatever happened later in the 2023–2025 unsecured cycle has to be measured against a known worst case, and the pandemic remains that worst case. Investors who lived through 2020 and 2021 watched this business absorb a shock far larger than anything the credit cycle subsequently produced, and emerge without a capital raise. That is a real, tested data point about the resilience of the model, and it deserves to be weighed alongside the criticisms that follow.

It is tempting to file this as bad luck, and in the narrow sense it was. But the more useful reading is that the IPO priced a specific claim — that Indian card penetration would compound, and that SBI Card's distribution would let it compound faster than the industry — at a valuation that assumed the claim would be validated quickly and without interruption. What the following six years actually delivered was one pandemic-driven credit shock, one regulatory capital shock, one industry-wide unsecured stress cycle, and a step-change in the cost of funds. The claim was not refuted. It was simply not paid for.

That distinction — a thesis that is intact but unpaid, versus a thesis that is broken — is the spine of everything that follows. And the first place to test it is the profit-and-loss statement itself, because a credit card company earns money in three quite different ways, and only one of them scales the way the IPO pitch implied.

↑ Back to Top

III. How a Credit Card Company Actually Makes Money

Picture a single SBI Card customer in Nagpur buying a washing machine for ₹40,000. Three separate revenue events fire at once, and they belong to three quite different businesses.

The first is a toll.

When the card is swiped, the merchant's bank pays a small slice of the transaction to SBI Card for having provided the credit and taken the fraud risk. This is interchange, and in SBI Card's accounts it shows up inside fees and commission income. It is beautiful revenue: it arrives whether or not the customer ever borrows, it scales directly with spending, and it carries no credit risk beyond the transaction itself. It is also low-margin per rupee and shared with the card network.

The second is a subscription and service business — annual fees, renewal fees, late-payment fees, EMI conversion charges, and the incentives paid by co-brand partners. This is why co-branded cards matter: an airline, a fuel retailer or an e-commerce platform pays for access to the cardholder, and SBI Card books it. The BPCL SBI Card, a fuel co-brand, crossed five million cards in the June 2026 quarter — a reminder that in India, where fuel is a weekly cash purchase for tens of millions of households, a petrol pump can be a better customer-acquisition channel than a shopping mall.⁴⁰

This fee line has become steadily more important. Fees and commission income rose 15% in FY26 to ₹9,166 crore, comfortably outgrowing interest income at 6%, and the split of revenue from operations in the March 2026 quarter was roughly 48% interest income and 52% everything else.⁵ That mix shift is double-edged. Fee revenue is capital-light and does not default. It is also more contestable, more visible to regulators — the RBI has repeatedly intervened on card fees and charges — and lower-margin per unit than lending.

The third is the lending business, and this is where the money actually is.

If the customer pays the full ₹40,000 by the due date, SBI Card earns the toll and nothing else. If the customer pays the minimum and revolves the balance, SBI Card earns interest at a rate that would be usurious in most contexts and is entirely ordinary in this one. If the customer converts the purchase to an EMI, SBI Card earns a lower rate over a longer, more predictable tenure.

The industry shorthand divides customers into transactors, who pay in full, and revolvers, who carry a balance. Transactors are the polite guests; revolvers pay for the party. A third category, the EMI customer, sits between them: lower yield than a revolver, but a scheduled, amortising, more predictable asset — closer in character to a personal loan than to a card balance. In SBI Card's disclosed receivables mix at March 2026, transactors accounted for 46% of receivables, EMI balances 32%, and revolvers just 22% — against 41%, 35% and 24% respectively a year earlier.⁵ That shift is the single most under-discussed fact about the company's recent recovery. The book got safer. It also got structurally less profitable per rupee: portfolio yield fell from 16.7% to 16.5% across FY26, and on the June 2026 quarter's call management told analysts it expected revolver balances to "continue to be stable where they are right now," with a slight downward bias.⁵¹⁰

Hold that thought, because it bounds the bull case from management's own mouth. The revolver pool is the highest-margin asset the company owns, and the people running it are guiding to flat-at-best.

Now the other half of the economics: the funding.

This is the part that makes SBI Card genuinely different from every competitor of comparable size, and it is worth explaining without jargon.

A bank that issues credit cards is funding those receivables with deposits. A salary account paying 3% interest, or a current account paying nothing at all, is raw material bought at a discount. The analogy is a restaurant that owns the farm: whatever happens to the price of vegetables in the market, its input cost barely moves.

SBI Card cannot gather deposits. It is a non-banking financial company, and it buys its raw material at wholesale prices that move every day — non-convertible debentures, commercial paper, working-capital demand loans and term loans from banks.⁵ Its cost of funds is therefore set by bond investors, by the policy rate, and by what rating committees think of it. Both CRISIL and ICRA rate it AAA/Stable on the long term and A1+ on the short term, and both explicitly lean on the expectation of support from State Bank of India in doing so.¹¹¹²

That rating is the single most valuable thing SBI's ownership provides — more valuable, arguably, than the branch network. It is also a borrowed asset rather than an owned one. A standalone monoline card lender in India with no parent would not carry AAA.

The numbers show both the benefit and the limit.

SBI Card's cost of funds fell from 7.5% in the March 2025 quarter to 6.6% in the March 2026 quarter, a 71-basis-point improvement across FY26 that flowed almost entirely to the margin line: net interest margin rose 31 basis points to 11.2% for the year even as asset yields fell.⁵ In other words, FY26's margin improvement was a gift from the rate cycle, not a product of pricing power. And by the June 2026 quarter the gift was being withdrawn — net interest margin fell 41 basis points year on year to 10.8%, and the company told analysts it expected the cost of funds to trend higher.⁶¹⁰

This is the central economic fact of the business, and it should be carried through every section that follows. Every bank-owned competitor funds the same receivable more cheaply than SBI Card can. The entire investment case rests on whether the distribution advantage is large enough, and durable enough, to pay for that gap. Which raises the obvious next question: how large is the gap, and who exactly is on the other side of it?

↑ Back to Top

IV. Industry Structure: Who Competes for India's Wallet

There is a statistic that has sold more Indian credit card equity stories than any other, and it deserves both its billing and its caveats. At April 2026, India had 119.44 million credit cards outstanding, growing 8.19% year on year.¹³ In a country of roughly 1.4 billion people, that is a card for something like one in twelve residents. In the United States, the great majority of adults carry at least one. The headroom is real and it is enormous.

Now the caveats, which matter more than the headline.

A "card in force" is not a customer — affluent Indians carry three or four, and the same wallet appears repeatedly in the national count. Growth is not linear: after the industry added cards at a furious pace through FY23 and FY24, issuance decelerated sharply through FY25 as lenders pulled back from unsecured credit. And penetration is not the binding constraint it once was, because UPI has already solved everyday payment for hundreds of millions of Indians without a card. What India is short of is not payment rails. It is formal, priced, revolving consumer credit — which is a narrower and more credit-sensitive market than the penetration statistic implies.

Within that market, four issuers own roughly seven of every ten cards. On the RBI's April 2026 data, HDFC Bank led with 26.44 million cards, SBI Card followed with 22.24 million, ICICI Bank held 19.20 million, and Axis Bank 16.09 million.¹³ Below them sit Kotak Mahindra, RBL, IDFC First and IndusInd, each in the low single digits of share. SBI Card is the only non-bank in the top tier — and the only issuer in that group whose entire business is this business. HDFC Bank can lose money on cards for two years and barely notice. SBI Card cannot.

The spending data from the same month tells a more interesting story than the card counts. HDFC Bank's cardholders spent ₹58,106 crore, up 12% year on year; SBI Card's spent ₹37,940 crore, up 29%; ICICI Bank's spent ₹32,499 crore, down 7%; Axis Bank's spent ₹22,023 crore, up 4%.¹³ SBI Card, in other words, has been winning the spends race decisively while its card-count growth trails HDFC's. That divergence is not an accident, and it is not entirely a compliment — a large slice of it is corporate and commercial card volume, which will matter a great deal in the next section.

Two structural events reshaped the competitive board in this period, and they cut in opposite directions.

Both are worth dwelling on, because they define the outer edges of what is possible in this industry: one shows what a well-capitalised bank can buy, and the other shows what a regulator can take away.

The first was consolidation.

Axis Bank announced the acquisition of Citibank's India consumer business in March 2022 and completed it on March 1, 2023, paying roughly $1.6 billion for a portfolio that included about 2.5 million affluent Citi cardholders along with deposits, wealth and mortgage relationships.¹⁴ Strategically it was the cleanest move anyone in the sector made in the decade: Axis bought, in one transaction, the premium customer cohort that takes a decade to build organically. It is also the natural benchmark against which SBI Card's own capital allocation should be judged, a comparison that returns later in this story.

The second was regulatory discipline applied to an incumbent.

On April 24, 2024, the RBI barred Kotak Mahindra Bank from onboarding new customers through online and mobile channels and from issuing fresh credit cards, citing deficiencies in IT risk management and repeated core-banking outages.¹⁵ The restrictions were lifted on February 12, 2025, after remedial work and an RBI-approved external audit.¹⁶ Ten months is a long time to be unable to acquire a customer. The episode is a reminder that in Indian financial services the regulator is a competitor's best friend and every incumbent's tail risk — a point that applies to SBI Card as squarely as it did to Kotak.

Run the industry through Porter's five forces and the picture sharpens. Rivalry among five or six similarly-scaled issuers is intense and largely fought on rewards, co-brands and acquisition spend rather than on price of credit. Buyer power is rising, because switching a card is nearly frictionless and reward-arbitrage behaviour among Indian consumers is now sophisticated. Barriers to entry are moderate — building a card business requires capital, a licence, and collections infrastructure — but new entrants keep arriving in adjacent form. And supplier power is the swing factor: for a bank issuer, the "supplier" of funds is its own depositor base, which is sticky and cheap; for SBI Card, the supplier is the bond market, which is neither.

Substitutes are where the analysis gets genuinely interesting, and here the regulator has been remarkably active. Four interventions between 2022 and 2026 each attacked a different piece of the business:

Tokenisation, effective October 1, 2022. Merchants, payment aggregators and e-commerce platforms were barred from storing actual card credentials, having to replace them with merchant-specific tokens; the deadline had already been extended twice from January 2022.¹⁷ The security logic was unimpeachable. The commercial effect was a one-off compliance cost across the industry and durable friction in the card-on-file and subscription flows that had been quietly compounding.

RuPay credit cards on UPI, live since 2022. A structural change important enough to warrant its own section below.

Risk weights, November 2023. The RBI raised the capital charge on unsecured consumer credit from 100% to 125% for banks and NBFCs alike, and raised it on credit card receivables specifically — to 150% for scheduled commercial banks and to 125% for NBFCs.¹⁸ For a diversified bank, this was a tax on one product line. For SBI Card, whose book is unsecured by definition and by construction, it was a tax on the entire company.

Card-network portability. The RBI's directions on card-network arrangements bar issuers from exclusive tie-ups with a single network, require customers to be offered a choice of network at issuance, and extend that choice to existing cardholders at renewal, with an exemption for issuers holding fewer than ten lakh active cards.¹⁹ Read as strategy rather than consumer protection, this is an assault on one of the few classic moat sources a card issuer has. Exclusive network economics — the volume deals, the co-marketing budgets, the bundled acceptance — were a genuine source of pricing power. They are now, by regulation, contestable.

The pattern across all four is worth naming: India's regulator has spent four years systematically converting the credit card business from a relationship business into a commodity utility with consumer-choice obligations attached. That is excellent public policy and a headwind for anyone claiming a durable moat. Which brings us to SBI Card's specific claim.

↑ Back to Top

V. The SBI Distribution Engine: Real Moat or Rented Advantage?

Walk into an SBI branch in a district town — Dholpur, say, or Warangal — and the credit card is not the product being sold. The savings account is. The card is what the relationship manager mentions afterwards, to a customer whose salary credit, balance history, loan repayment record and KYC file the bank has already held for years. The customer is pre-scored before the conversation begins. Acquisition cost is a fraction of what it costs a fintech to buy the same person's attention on a performance-marketing platform, and the credit information is better than any bureau file.

That is the moat claim, and it is not a trivial one. As of March 2026, 54% of SBI Card's new account sourcing came through the SBI channel, with 46% from the open market; on the total cards-in-force base the split was 64% SBI-sourced to 36% open market.⁵ No standalone competitor has anything remotely like it, and no bank competitor has a branch network of comparable reach into tier-2 and tier-3 India — 30% of SBI Card's new sourcing in the March 2026 quarter came from tier-2 cities and 20% from tier-3, with only 30% from tier-1.⁵

There is a second, subtler advantage layered on top of the first, and it is worth naming because it is easy to miss. Sourcing through the parent's own customer base is not just cheaper — it is better informed. A bank knows things about its customer that a credit bureau does not: the volatility of the salary credit, the pattern of balance drawdown before month-end, whether the EMI on the home loan is ever paid a day late. An open-market applicant arrives with a bureau score and a set of documents. An SBI-sourced applicant arrives with a decade of behavioural history attached. In theory, that should produce measurably better credit outcomes on the SBI-sourced cohort.

Does it? SBI Card's own disclosure offers a partial answer. In its March 2026 portfolio breakdown, the company indexed 30-plus-day delinquency across its cards-in-force by sourcing channel, and the open-market cohort indexed worse than the SBI-sourced cohort.⁵ The direction is what the theory predicts. The magnitude, though, is not dramatic, and the company does not disclose the underlying rates — only the index. So the informational advantage appears to be real but modest, which is roughly what one should expect: bureau data in India has improved enormously over fifteen years, and the proprietary edge a bank once held over an outside lender has narrowed accordingly.

So far, so good for the bulls. Now the disconfirming evidence, which belongs here rather than in a distant risk list.

The channel did not prevent share drift during the stress cycle.

Through 2024, as SBI Card tightened underwriting, its cards-in-force share slipped while bank-owned rivals kept issuing.²⁰ That is the test case that matters: the distribution advantage was fully intact and it did not insulate the company when the economics of the marginal card turned. Distribution determines who can be offered a card cheaply. It does not determine who wants to issue one.

Card growth still lags the industry leader. In the twelve months to April 2026, SBI Card added 1.25 million cards. HDFC Bank added 2.44 million.¹³ SBI Card's own base grew 6% in FY26 against industry growth of about 8%.⁵¹³ The company can point to the June 2026 quarter, where it reported the highest net card additions in the industry at 484,000 and new accounts up 17% year on year, and that is a genuine inflection worth crediting.⁶ But one quarter of leadership after two years of trailing is a data point, not yet a trend.

The spends share is flattered by low-margin volume.

This is the most important qualification, and it is visible only if you look past the headline. SBI Card's total spends grew 29% in FY26 to ₹430,359 crore. Retail spends grew 13%. Corporate spends grew 195% year on year in the March 2026 quarter alone, from ₹8,656 crore to ₹25,564 crore.⁵ Corporate and commercial card volume is real business, but it is thin-margin transaction processing: it almost never revolves, it earns little interchange relative to retail, and it is priced competitively because corporate treasurers are professional buyers. Meanwhile receivables — the actual lending book, the thing that generates interest income — grew 2% for the year, from ₹55,840 crore to ₹56,926 crore, and receivables per card fell from ₹26,816 to ₹25,762.⁵

Put those together and the picture is unambiguous. SBI Card's spends market share reached 19.5% in the June 2026 quarter against cards-in-force share of 18.6%, and management is entitled to present that as leadership.⁶ But a company whose spends are compounding at 29% while its loan book compounds at 2% is not converting distribution into earning assets. It is converting distribution into payment volume, which is a materially less valuable thing.

There is one more piece of counterevidence that deserves airing, because it goes to the durability rather than the size of the advantage. The SBI channel is not contracted, priced, or owned. It exists because the promoter chooses to make it available. Nothing in the public record establishes what SBI Card pays for that access, how the arrangement is renewed, or what would happen if the parent decided to build out its own in-house card issuance more aggressively. For a minority shareholder, the single most valuable asset in the investment case is a related-party arrangement whose terms are not disclosed in a form that permits independent valuation. That is not a scandal — it is entirely ordinary in Indian promoter-subsidiary structures — but it is a real limit on how confidently anyone outside the company can call this a moat rather than a courtesy.

So what is the verdict on the moat claim? The historical record does not reject it — the SBI channel is demonstrably real, demonstrably cheap, and demonstrably reaches parts of India that rivals do not. But it narrows the claim substantially. What the evidence supports is a sourcing cost advantage, not a customer retention or pricing advantage. Nothing in the record shows SBI Card earning better unit economics than bank-owned peers on a comparable customer; what it shows is that the company can find that customer more cheaply and then must fund the resulting receivable more expensively.

The KPI that would resolve this is specific: whether SBI Card's cards-in-force share and — more importantly — its receivables growth re-accelerate relative to HDFC, ICICI and Axis through FY27 and FY28, now that the credit cycle has turned and underwriting has loosened. If spends keep compounding while receivables stay flat, the distribution engine is producing volume rather than value, and the moat claim should be downgraded to "sourcing advantage, roughly offset by funding-cost disadvantage."

Distribution, though, is only as good as the people deciding what to do with it. And on that front SBI Card has a specific and unusual arrangement.

↑ Back to Top

VI. Current Management, Ownership, and Capital Allocation

On April 1, 2025, Salila Pande took over as Managing Director and Chief Executive Officer of SBI Card, on a two-year term.²¹ Her CV is instructive. She joined State Bank of India as a probationary officer in 1995 and spent almost three decades there — most recently as Chief General Manager of the Mumbai Metro Circle, running SBI's retail banking across India's most important commercial market, and before that as President and CEO of SBI California, where she managed a small overseas subsidiary through the pandemic. She holds a postgraduate degree in physics and is a certified Financial Risk Manager.²¹ SBI's chairman, Challa Sreenivasulu Setty — who also chairs SBI Card's board as non-executive chairman — framed her appointment around "new product development, innovation, expansion and value creation."²¹²²

She is the third SBI career banker to hold the job in five years. Rama Mohan Rao Amara led the company through the pandemic recovery. Abhijit Chakravorty — an SBI man since 1988 who had served as CEO of SBI's Bangladesh operations and as a Deputy Managing Director at the parent — took charge in August 2023, ran the company through the worst of the credit cycle, and retired on superannuation on March 31, 2025.[^23]²¹ The Chief Financial Officer, Rashmi Mohanty, was appointed with effect from October 21, 2022, and brings a fixed-income, foreign-exchange and treasury background — a useful skill set at a company whose margins are set in the bond market.²³²²

Chakravorty is worth a moment on his own, because his tenure maps exactly onto the hardest two years in the company's listed history. He had joined State Bank of India in 1988 and spent thirty-five years inside it, including a posting as chief executive of SBI's Bangladesh operations and a stint as a Deputy Managing Director at the parent — a career built on running large, distributed banking operations rather than on consumer credit modelling.[^23] He arrived in August 2023, three months before the RBI's risk-weight circular, and inherited a book that was already deteriorating without anyone yet knowing how far it would go. His public style on earnings calls was patient and specific: he listed remediation actions in detail, took analyst pushback without deflecting to macro conditions, and never blamed his predecessor. He also, as the transcripts show, told the market the peak was near before it was.

The pattern is the point. SBI Card's leadership is an SBI posting, not an independently recruited management team, and the terms are short — two years, tied to the parent's superannuation calendar. There are real advantages: the person running the subsidiary has personally run pieces of the distribution channel it depends on, and the alignment with the promoter is total. There are equally real costs. A two-year term is shorter than a credit cycle. A CEO who arrives in year one of a stress cycle and departs before it resolves cannot be held accountable for the outcome in any meaningful sense — and, more subtly, has little incentive to take the kind of multi-year strategic risk that would change the company's structural position. Chakravorty's tenure is the case in point: he inherited a deteriorating book, told the market repeatedly that the peak was near, and left before the improvement he had predicted actually arrived.

Ownership reinforces the same structure.

At March 31, 2026, the SBI promoter group held 68.92%, with State Bank of India itself at 68.58%. Mutual funds held 10.65%, foreign portfolio investors 9.54%, insurance companies 6.55% and resident individuals 2.87%.⁵ The promoter stake is not going anywhere; SBI has repeatedly signalled that the card business is core. The practical consequence for a minority shareholder is that the free float is small relative to the market capitalisation, the shareholder register is dominated by domestic institutions, and there is no realistic path by which an outside investor influences strategy.

There used to be one such investor. Carlyle, holding through CA Rover Holdings, sold down its post-IPO position in a series of block trades: a partial exit at the IPO itself, then roughly 4.3% for about ₹3,944 crore in March 2021, a further tranche, a 3.4% block worth roughly $443 million, and a final 2.78% for about ₹2,229 crore in April 2022.^24252627 Carlyle's own portfolio page now lists the SBI Card investment as "Exited."⁹

It is worth being precise about what this signals and what it does not. A private equity fund exiting within roughly two years of a listing is not, in itself, evidence that the business is bad — that is what growth-equity funds do, and Carlyle had held the asset since December 2017. But a full exit is a materially stronger signal than a trim, and the timing matters: Carlyle walked away completely in April 2022, well before the credit cycle turned, at a point when the growth story was still intact and the shares were well above where they trade today. What is unambiguous is the governance consequence. The register no longer contains a single large holder with an independent, non-SBI commercial incentive. The "sophisticated sponsor validates the story" argument, which was part of the IPO pitch, does not apply to today's shareholder base.

Now to capital allocation, where the record deserves a harder look than it usually gets.

No acquisition appears in SBI Card's record — not one, across twenty-eight years; growth has been entirely organic.

Set against Axis Bank's Citi transaction, which reshaped that bank's card franchise in a single stroke, the contrast is stark: the company with the largest distribution funnel in Indian retail finance watched a portfolio of 2.5 million affluent, high-spending Indian cardholders — precisely the cohort where its own book is weakest — change hands, and did not bid.

It is worth being fair about why. SBI Card is an NBFC subsidiary of a public-sector bank, and any acquisition of scale would have required promoter approval, regulatory clearance, and — because the company generates capital rather than holding surplus cash — either debt or equity issuance at a moment when the RBI was actively discouraging unsecured growth. The window in which the Citi portfolio was available was also, in hindsight, a window in which SBI Card's own credit quality was about to turn. A bid would have been aggressive.

There are two readings. The charitable one is discipline: SBI Card did not overpay for a portfolio at the top of a cycle, and it has no history of value-destroying deals to explain away. The uncharitable one is that "never having done a deal" is not evidence of acquisition discipline, because discipline requires having faced the choice and declined it for articulable reasons. Nothing in the public record establishes which of these is true. What can be said is that the company has never been tested on capital allocation in the way most listed financials eventually are, and that a claim of "disciplined capital allocator" rests on a thin base: a steady dividend of ₹2.50 per share declared for FY26 against basic earnings of ₹22.77 — a payout ratio of roughly 11% — and the absence of destructive M&A.²⁸⁵ That is a low bar cleared, not a skill demonstrated.

The capital itself has been managed conservatively, and here the record is better. Through the worst of the stress cycle, capital adequacy never approached the regulatory floor: the ratio bottomed around 20.6% in the June 2024 quarter and had rebuilt to 25.5% by March 2026, with Tier 1 at 20.0%.²⁹⁵ No dilutive equity raise was required at the bottom of the cycle — a genuinely meaningful fact, because forced issuance at a depressed price is how monoline lenders permanently destroy shareholder value. Total equity grew from ₹13,782 crore to ₹15,725 crore across FY26 through retained earnings alone.⁵

One caveat on disclosure. Incentive structures and employee stock option details were not independently verifiable from the interim filings reviewed here; a reader assessing management alignment should go directly to the compensation and ESOP tables in the latest annual report rather than assume it.

Which brings us to the period that actually tested all of this.

↑ Back to Top

VII. The Reckoning: Risk Weights, Rising Delinquencies, and the Credibility Test (2023–2026)

On November 16, 2023, the Reserve Bank of India did something that changed the arithmetic of an entire industry overnight. It raised the risk weight on unsecured consumer credit from 100% to 125% for banks and NBFCs, and raised the weight on credit card receivables specifically — to 150% for scheduled commercial banks and 125% for NBFCs.¹⁸

Understand what a risk weight does, because it is the most powerful lever in financial-services regulation and the least intuitive.

Capital adequacy is calculated as equity divided by risk-weighted assets. Raising the risk weight on a loan does not change the loan; it changes how much shareholder equity must sit behind it. A 25-percentage-point increase means that for every ₹100 of card receivables, the lender must now hold capital as though it had lent ₹125. Returns on equity fall mechanically, before a single borrower has missed a payment.

It is a tax on growth in one specific product, imposed instantly, and there is no appeal.

For a diversified bank, this trims one line of a large book. For SBI Card, whose receivables are unsecured by definition, it was a levy on the whole enterprise. The RBI's intent was explicit: unsecured retail credit had been growing at multiples of overall credit growth, and the central bank wanted the brakes applied before the cycle broke. The brakes worked. Industry-wide issuance decelerated sharply through FY25, and SBI Card's own new account additions fell from 4.36 million in FY24 to 4.09 million in FY25 and then to 3.59 million in FY26.²⁹⁵

But the regulator was reacting to something already underway, and here the company's numbers tell the story better than any commentary. Gross NPAs, which had been 2.14% in the September 2022 quarter, reached 2.76% at the close of FY24 and 3.08% at the close of FY25. Gross credit loss — the cost of writing off bad receivables, expressed as a percentage of the book — climbed from 7.1% in FY24 to 9.0% in FY25. Net profit fell 20%, from ₹2,408 crore to ₹1,916 crore. Return on equity dropped from 21.7% to 14.6%.²⁹

For useful perspective: this was not the worst credit event in the company's listed history. The pandemic peak, described earlier, was materially higher. But it was also a single, visible, exogenous shock that arrived and departed. The 2023–2025 cycle peaked lower and lasted longer. It was a grind, not a shock — no lockdown, no moratorium, no headline event, just delinquency rates that would not stop drifting upward quarter after quarter.

That difference is precisely why it was harder to forecast, and why management's real-time commentary is worth examining closely. A shock forces humility; a grind invites the belief, renewed each quarter, that the next print will be the one that turns.

The record, quarter by quarter.

On the Q3 FY24 call in January 2024, with credit cost stepping up sequentially, management framed the deterioration as portfolio normalisation after an unusually benign pandemic period, attributed the stress to identifiable geographies, and emphasised proactive portfolio actions taken ahead of accounts reaching NPA status.³⁰ Analysts pushed back, questioning whether this represented a sustainable level or a temporary spike, and specific guidance on where credit cost would stabilise was not forthcoming. Brokerages downgraded the stock on the print.³¹

By the Q1 FY25 call in July 2024, the explanation had shifted — and this is where it becomes analytically interesting. Chakravorty told analysts that credit cost had risen to 8.5% from 7.5% in the prior quarter, and gave a new causal account: "customers obtained multiple trade lines from other lenders after taking a card, and this overleveraging has impacted their repayment capacity." He added that reduced payment capacity was showing up even "with vintage customers having good repayment behaviour until now," and listed the interventions — limit reductions across five lakh accounts in three months, cross-sell restrictions, spend-trigger-based early blocking, scorecard refinement, expanded collections capacity. Asked directly about the prior guidance, he was specific: "We had given a guidance of upwards of 7%. So, we do see a downward trend during the latter part of the year. And as of now we can only assume it to remain between 7% and 8%. We will continue to hold on to that to be between 7% and 8%."³²

Gross credit loss for FY25 came in at 9.0%.²⁹

By the Q2 FY25 call on October 29, 2024, with credit cost at 9%, the language had moved to peak-calling: "We believe we are closer to the peak. Our flows into delinquency have improved over the last six months, this gives us some confidence about the efficacy of our actions." Pressed by an analyst from Nuvama on whether peaking flows meant peaking credit cost, Chakravorty was careful but held the line: "What we stated just now is that we see ourselves closer to the peak. It will take a couple of quarters more, at least one or two quarters more to understand the pattern."³³

Credit cost rose again in the following quarter, to roughly 9.4%.³⁴

Three observations follow from this sequence, and they should be weighed rather than simply listed.

First, the misses were real and repeated.

A full-year guidance range of 7–8% delivered at 9.0% is not a rounding error; it is a 100-plus-basis-point overshoot on the single most important cost line in the business, given roughly nine months into the year. A "closer to the peak" call in October 2024 was followed by a further deterioration in January 2025. This was not one unlucky quarter — it was a pattern of forecasts overtaken by events across roughly six quarters.

Second, the explanations shifted without ever fully resolving.

In January 2024 the cause was normalisation and city-specific stress. By July 2024 it was borrower over-leverage across multiple lenders. By October 2024 it was cash-flow stress and reduced willingness to cure once delinquent. Each explanation was plausible; none was retracted; and the company never offered a single coherent account of what had actually gone wrong in its own underwriting as distinct from the industry's. When a lender's diagnosis migrates across three distinct causes in three quarters, the honest reading is that it did not know, in real time, what it was looking at.

Third — and this cuts the other way — the same management team was simultaneously right about something important. On that October 2024 call, management also stated that the cost of funds "has peaked out and will start coming" down.³³ It had, and it did: funding costs fell 71 basis points across FY26.⁵ The forecasting weakness was specific to credit, not general.

And the actions, whatever the commentary, worked.

FY26 delivered the turn: gross NPAs fell 67 basis points to 2.41%, net NPAs to 1.04%, gross credit loss to 8.6% and net credit cost to 7.4%. Profit after tax rose 13% to ₹2,167 crore. Provision coverage improved to 57.6%, and the company carried an additional provision of ₹220 crore as of March 2026 — a conservatism buffer above the modelled expected-credit-loss requirement.⁵ The June 2026 quarter extended the improvement sharply: gross credit cost at 6.5%, down 301 basis points year on year and 116 basis points sequentially; net NPAs at 0.83%, the lowest since the December 2022 quarter; profit up 20% to ₹664 crore.⁶¹⁰

Here, though, is the qualification that a numbers-only reading would miss.

FY26's profit recovery was almost entirely a credit-cost story, not an operating one. Total income rose 11% to ₹20,708 crore, but earnings before credit cost rose only 6%, to ₹7,876 crore, because operating costs jumped 22% to ₹9,760 crore and the cost-to-income ratio deteriorated by 355 basis points to 55.3% — reaching 57.2% in the March 2026 quarter.⁵ Management has guided to 55–58% as the steady-state range.³⁵¹⁰ So the company earned its way back to growth by making fewer bad loans, not by running the business more efficiently. That is a legitimate and valuable form of recovery. It is not the same as operating leverage, and it does not repeat.

Net assessment.

The evidence narrows the "disciplined, well-managed growth story" claim without rejecting it. The disconfirming facts are specific and documented: guidance missed by a wide margin, a peak called at least twice before it arrived, and diagnoses that migrated quarter to quarter. The mitigating facts are equally specific: the peak of this cycle — gross NPAs of 3.27% in the September 2024 quarter — was well below the pandemic peak;³⁴ no capital raise was forced; the portfolio actions taken from mid-2024 demonstrably worked, with FY26 and the June 2026 quarter as the proof; and the industry as a whole, including bank-owned issuers with far more diversified data, misjudged the same cycle.

The revised claim that the record supports is modest: SBI Card's management executes competently on portfolio remediation once stress is identified, and forecasts it poorly in advance. The forward test is precise. Watch whether gross credit cost and gross NPA keep improving through FY27 and FY28 without another guidance reversal. A third "near the peak" call that proves wrong would be a far stronger signal against management credibility than this cycle on its own.

One live regulatory thread belongs here rather than in a generic risk list. On October 8, 2025, the RBI issued a draft circular on credit-risk capital that proposed reducing the risk weight on transactor credit card exposures — customers who repay in full every month — from 125% to 100%, while holding non-transactors at 125%. Analysts calculated that with roughly 40% of SBI Card's then-₹56,607 crore receivable book sitting in the transactor bucket, the change could release capital equivalent to around 450 basis points; the shares rose on the news.³⁶ The final rules — the Reserve Bank of India (Commercial Banks – Capital Charge for Credit Risk – Standardised Approach) Directions, 2026 — were issued on April 27, 2026, effective April 1, 2027, and place qualifying transactor exposures in the 75% regulatory retail portfolio while leaving other credit card receivables at 125%.³⁷ The critical caveat, which the market commentary tends to skip: those Directions govern commercial banks. SBI Card is an NBFC. Whether an equivalent recalibration is extended to non-bank issuers is not settled, and it is one of the more consequential open regulatory questions for this specific company — because, as noted, its transactor share has been rising.

↑ Back to Top

VIII. The RuPay-UPI Question: Disruption or Distribution Channel?

Consider a vegetable seller in a tier-3 town who has never owned a card machine and never will. Two years ago, a customer wanting to pay by credit had no way to do so. Today that customer opens a UPI app, scans a QR code taped to a crate, and pays from a RuPay credit card. No terminal, no merchant discount rate negotiation, no acquiring bank relationship. The credit line has been unbundled from the plastic and from the point-of-sale hardware entirely.

This is the single most consequential structural change in Indian payments since UPI itself, and it is genuinely unresolved which way it cuts.

Start with the plumbing, because the economics follow from it. A conventional card transaction runs across a four-party system: the cardholder, the issuing bank, the merchant's acquiring bank, and a network — Visa, Mastercard, RuPay or Diners — that sets the rules and clears the transaction. The merchant pays a discount rate, most of which flows back to the issuer as interchange. That interchange is what funds reward points, lounge access, cashback and the acquisition budgets that Indian issuers have spent a decade escalating. Credit on UPI keeps the four-party structure but changes two of its parameters: the network is RuPay, and the merchant-side pricing is governed by a domestic regime in which small-merchant transactions have historically been priced far more thinly than card-present retail.⁴¹

The threat case is straightforward.

The economics of a premium credit card rest on interchange, which rests on merchant discount rates, which rest on a card-acceptance infrastructure that Visa and Mastercard built and monetise. Credit on UPI runs on RuPay rails, is subject to a domestic pricing regime rather than an international one, and is being deliberately promoted by policy. By early 2025, roughly 16% of all Indian card spending was happening on RuPay, and nearly half of that was credit routed through UPI; average monthly UPI-based RuPay spending had risen to around ₹18,000 crore from about ₹10,000 crore eighteen months earlier, with more than thirty banks issuing.³⁸ Transaction volumes for RuPay credit on UPI reached over 750 million transactions worth ₹63,826 crore in the first seven months of FY25, against 362.8 million transactions worth ₹33,439 crore in all of FY24.³⁸ If that trajectory continues and the associated interchange economics remain thinner than card-present or card-not-present norms, network revenue per rupee of spend compresses across the entire industry.

The opportunity case is equally coherent, and it is specific to SBI Card's position.

The company's customer base skews mass-market and geographically dispersed in a way HDFC's and Axis's premium books do not — 50% of its cards-in-force sit outside tier-1 cities.⁵ That is exactly the population for whom UPI is the default payment method and for whom card acceptance was historically the binding constraint. SBI Card's own disclosures make the point vividly: among its UPI-active cardholders at March 2026, 77% sat in tier-2 and beyond, and those customers accounted for 81% of UPI spends, with an average monthly UPI spend per account of ₹22,400 against ₹17,700 for tier-1 users.⁵ The top spend categories were departmental stores and grocery, fuel, utilities, apparel and restaurants — everyday, high-frequency, low-ticket transactions that were simply not addressable by a card before.

In other words, the same mechanism that threatens premium interchange pools may be opening an entirely new addressable market at the bottom of SBI Card's own base. UPI spends on the company's RuPay cards grew more than 10% sequentially in the March 2026 quarter alone.⁵ And 16.89% of the credit cards SBI Card issued in FY26 were RuPay cards enabled for UPI.⁵

There is a third possibility that sits between the two, and it may be the likeliest. Credit on UPI may function less as a spending product and more as an engagement product — the thing that keeps a card top-of-wallet for a customer who would otherwise use it twice a month. Engagement is not directly monetisable, but in a lending business it is the precondition for everything that is: a card that is used weekly generates data, gets a limit increase, becomes eligible for EMI conversion, and eventually revolves. On that reading, the low interchange on a ₹200 grocery transaction is a customer-acquisition cost paid in foregone margin rather than in marketing spend. Whether the downstream conversion actually happens is an empirical question that neither SBI Card nor any peer has yet disclosed data on.

The honest position is that this is unresolved. Three things would need to be observable before anyone can call it: whether the interchange economics of credit-on-UPI transactions settle at a level that supports the rewards and acquisition spending the industry currently runs; whether the incremental UPI user revolves or converts to EMI at rates comparable to a conventional cardholder, or simply transacts; and whether the volume is genuinely incremental or is cannibalising card-present spend the company was already capturing. None of these has been disclosed with enough granularity by anyone to settle the question.

What can be said is that SBI Card is positioned better for this transition than a premium-skewed issuer would be, and that the company is leaning into it. Whether that turns into earnings depends on economics that have not yet been published. A new distribution rail is not revenue until it is priced.

↑ Back to Top

IX. Playbook: Business & Investing Lessons

Strip away the specifics and a handful of transferable lessons sit inside this story. None of them is unique to Indian credit cards; all of them are unusually clearly illustrated here, because SBI Card is close to a laboratory specimen — one product, one customer type, one funding structure, one promoter, and six years of public data covering two distinct credit shocks.

A distribution advantage is only as good as the cost of capital underneath it.

SBI Card is close to a controlled experiment. Hold the customer-acquisition channel superior — genuinely, measurably superior, with 64% of the card base sourced from the parent's branch and relationship network — and make the funding structurally more expensive than every peer's. What happens? The answer, over six years, is that the two roughly cancel. Returns on equity land in the mid-teens rather than the low twenties a bank issuer with the same book would produce, and the share price goes nowhere. For investors, the generalisation is that in any spread business — lending, insurance, leasing, asset management — a demonstrable advantage on the revenue side is routinely neutralised by a structural disadvantage on the cost-of-funds side, and the second is far harder to fix than the first. Cheap distribution can be built. Cheap deposits require a banking licence.

A full sponsor exit is a different signal from a partial trim, and it deserves more weight than the retention of a strategic name.

Carlyle sold everything, in tranches, within roughly two years of the listing. SBI stayed. It would be easy to read the second fact as reassurance and ignore the first. But the two shareholders had entirely different objectives: SBI holds for strategic and franchise reasons, and would hold at almost any valuation; Carlyle held for return, and stopped. When the only shareholder whose incentive is purely financial exits completely, that is information — not proof of anything, but information that a strategic holder's continued presence does not offset.

"Near the peak" is a phrase to be tracked across calls, not accepted within one.

The transcripts are unambiguous: sincere, specific, internally consistent reassurance was delivered in July 2024 and again in October 2024, and both were overtaken by the following two quarters. Nobody was being deceptive. Everybody was wrong. The practical technique this suggests is mechanical — keep a running log of every quantitative forward statement management makes, and mark it to actuals four quarters later. The gap between the two is a better measure of management quality than the eloquence of any single call. It is also worth noting the asymmetry: the same team's call on the cost of funds peaking proved correct. Forecasting skill is domain-specific, not general.

A monoline is a leveraged bet on one product cycle, and that cuts both ways.

The reason SBI Card's profit fell 20% in a single year is not that its underwriting was uniquely bad — bank issuers saw the same deterioration in their card books. It is that a bank's card book is 5% of its assets and SBI Card's card book is 100% of its assets. The same concentration is what produces returns on equity in the high teens to low twenties when the cycle is benign, and it is why the stock has been so much more volatile than the underlying business. Investors in single-product financials should size the position for the amplitude, not the average.

Regulatory capital rules are the most powerful lever in financial-services investing, and they operate independently of anything the company does.

A single November 2023 circular reduced the return on equity of the entire unsecured lending industry, throttled issuance growth, and did so with no reference to any individual lender's underwriting quality. Two years later, a draft circular moving transactor exposures to a lower weight moved the same stocks in the opposite direction. For anyone investing in regulated lenders, the risk-weight and provisioning framework is not a footnote in the risk section — it is a primary driver of returns, and it should be monitored with the same attention as credit quality itself.

Headline metrics and value metrics diverge, and companies will report the flattering one.

"Record spends of ₹1.18 lakh crore" is a true statement that SBI Card made about the June 2026 quarter, and it is the number that led the coverage.⁶ "Receivables grew 3%" is also true, less prominent, and far more informative about the earnings power of the business. This is not deception; every management team leads with its best line. The transferable habit is to identify, for any business, which single disclosed metric most closely tracks the economics — for a card issuer it is the interest-earning receivable, for a retailer it is same-store sales, for a subscription business it is net revenue retention — and to read every quarterly release starting from that number rather than from the headline.

Together these produce a specific way of framing the investment question, which is where the story now goes.

↑ Back to Top

X. Bull vs. Bear: The Investment Case

The bull case, stated at its strongest.

India's credit card market remains structurally under-penetrated at roughly 119 million cards for a population of 1.4 billion, and formal revolving consumer credit is scarcer still.¹³ SBI Card holds the number two position in cards and the number one or two position in spends, with spends share reaching 19.5% against cards share of 18.6% in the June 2026 quarter — meaning its customers spend more per card than the industry average.⁶ The credit cycle has demonstrably turned: gross NPAs have fallen for five consecutive quarters to 2.04%, net NPAs to 0.83%, and gross credit cost to 6.5%, all with a ₹220 crore provisioning buffer still held above the model requirement.⁵⁶ Capital adequacy at 25.6% is far above any binding constraint, meaning growth requires no equity issuance.⁶ The AAA/A1+ ratings from CRISIL and ICRA, resting on SBI's implicit support, keep funding costs meaningfully below what a standalone monoline would pay.¹¹¹² Card sourcing is re-accelerating, with June-quarter new accounts up 17% and the highest net additions in the industry.⁶ The dividend has been paid consistently. And management expects asset growth to pick up from the second half of FY27 on the back of the higher acquisition run rate.¹⁰

The bear case, stated at its strongest.

Start with the outcome: more than six years after listing, the stock trades roughly 14% below its offer price.⁷¹ That is not a rocky start — it is a full cycle of buy-and-hold underperformance, through a period in which the Indian market compounded substantially. The funding-cost gap versus bank-owned rivals is structural, permanent under the current corporate form, and does not improve when credit improves; management itself told analysts in June 2026 that funding costs would trend higher from here.¹⁰ The composition of the recovery is unflattering: profits rose in FY26 because credit losses fell, while operating costs rose 22% and the cost-to-income ratio deteriorated by 355 basis points.⁵ The lending book — the source of the high-margin interest income — grew 2% in FY26 while spends grew 29%, meaning the company is increasingly a payments processor and increasingly less a lender.⁵ The revolver share of receivables has fallen from 24% to 22% and management guides it flat with a downward bias, which caps the highest-yielding pool.⁵¹⁰ Management's forward commentary through the stress cycle was repeatedly wrong. Carlyle's full exit removed the only shareholder with an independent commercial incentive. And regulation has proven twice in three years that it can reshape the growth and return profile of this exact business without warning.

Myth versus reality, briefly. Three consensus statements about this company do not survive contact with the disclosures. The first is that SBI Card is "a play on Indian credit card penetration" — it is more accurately a play on Indian revolving credit demand, and those are different markets moving at different speeds; penetration has kept rising while the company's own loan book has been flat for a year. The second is that "SBI parentage means low funding costs" — parentage means an investment-grade rating that a standalone monoline could not achieve, which is not the same as bank-equivalent funding; the residual gap versus a deposit-funded issuer is the whole problem. The third is that the March 2020 listing was "unlucky timing" — the timing was unlucky, but six years of flat-to-negative returns is not a timing story, it is the market repricing a growth multiple to reflect a cyclical, capital-intensive, regulator-exposed lender.

Where the evidence is genuinely thin. The proposition that the distribution advantage is durable enough to offset the funding gap over a multi-year horizon remains unproven in either direction. Cards-in-force share data through the stress cycle is mixed rather than supportive: SBI Card grew its base 6% in FY26 against industry growth of about 8%, but posted the highest net additions in the industry in the June 2026 quarter.⁵¹³⁶ One quarter does not establish a trend and two years of underperformance does not establish a structural deficiency. This is a genuinely open question, and anyone who tells you otherwise is extrapolating.

How it compares. The cleanest way to frame the competitive position is to ask what SBI Card would look like if it were a division rather than a company. Inside HDFC Bank, the same book would be funded with deposits, would carry a 150% risk weight rather than 125% under the current NBFC rules, would cross-sell into mortgages and wealth, and would be judged on contribution rather than on standalone return on equity.¹⁸ Inside Axis Bank, it would sit alongside an acquired premium portfolio and a much larger corporate relationship base. Standalone, SBI Card gets a cleaner story for investors, a dedicated management team, and full accountability — and pays for all three in the cost of funds. Whether that trade was ever worth making for minority shareholders is the structural question that has hung over the stock since listing, and nothing in six years of results has resolved it in the affirmative.

Through Helmer's 7 Powers, the picture is sobering. Scale economies: partially present — SBI Card is large enough to spread fixed technology and collections costs, but its cost-to-income ratio at 55.3% is not evidence of scale advantage.⁵ Network economies: absent to the issuer; they accrue to Visa, Mastercard and RuPay, not to the card issuer, and network portability rules have made even the issuer's ability to monetise a network relationship contestable.¹⁹ Counter-positioning: absent — SBI Card does nothing that a bank issuer cannot copy, and the reverse is not true. Switching costs: weak and weakening; a cardholder can and does hold four cards, and the regulator has actively reduced friction. Branding: this is the real one. The SBI name carries trust in a country where trust in financial institutions is unevenly distributed, and it is the reason a tier-3 customer opens the envelope. Cornered resource: the closest thing SBI Card has — exclusive preferential access to the parent's customer list and branch network. But it is licensed rather than owned, and its value has been demonstrated in sourcing cost rather than in retention or pricing. Process power: not demonstrated; if anything, the 2023–2025 cycle is evidence against superior credit process relative to peers.

That leaves one and a half powers, both of which trace back to the parent rather than to anything the company built. This is what "rented advantage" means in practice, and it is why the funding-cost handicap matters so much: SBI Card does not have an offsetting structural advantage of its own to deploy against it.

The why-win / why-not test, stated plainly. SBI Card wins from here if India's under-penetration lets it convert its cheapest-in-industry sourcing into a growing, revolving, interest-earning book faster than funding costs and regulatory capital charges erode the spread — and if the RuPay-UPI transition turns into incremental mass-market credit demand rather than commoditised transaction volume. It does not win if the pattern of the last two years persists: spends compounding while receivables stagnate, transactor share rising while revolver share falls, operating costs growing faster than income, and bank-owned rivals with cheaper deposits out-executing on digital acquisition even with SBI Card's channel advantage fully intact.

An activist looking at this company would ask three uncomfortable questions. Why does a business with 68.58% promoter ownership and a AAA rating pay out only 11% of earnings when it is generating capital faster than it can deploy it and running capital adequacy at 25.5%?⁵²⁸ Why has a two-year CEO term been treated as adequate governance for a business whose credit cycles run four to six years? And why, given the parent's stated commitment to the card business as core, has the corporate structure — an NBFC funded wholesale, sitting beside a bank with ₹61 trillion of assets and a vast deposit base — never been revisited, when that structure is the single largest identifiable drag on returns?³ None of those questions has a published answer.

↑ Back to Top

XI. Risk Radar

Regulatory and capital risk sits at the top, and it is not symmetric.

The November 2023 experience established the mechanism: a single circular can raise the capital cost of SBI Card's entire book while touching only a slice of a diversified bank's. The forward version of this risk is specific and currently live — the RBI's 2026 credit-risk Directions extend favourable transactor treatment to commercial banks from April 2027, and whether an equivalent recalibration reaches NBFC issuers is undecided.³⁷ If bank issuers receive capital relief on their transactor books and SBI Card does not, a regulatory gap opens on top of the existing funding gap, in the very segment where SBI Card's mix has been shifting. This is the asymmetry to watch.

Refinancing and cost-of-capital risk is the permanent condition of the business, not a periodic event.

With borrowings of ₹44,064 crore at March 2026, funded through non-convertible debentures, commercial paper, working-capital demand loans and term loans, SBI Card reprices a material portion of its liabilities continuously.⁵ FY26's margin expansion came from a falling rate cycle; management has already flagged that funding costs will trend higher.¹⁰ The mechanism is direct and quick: every 100 basis points of funding cost on a ₹44,000 crore borrowing base is roughly ₹440 crore of pre-tax earnings, against FY26 profit before tax of ₹2,913 crore.⁵ A bank issuer's deposit base simply does not reprice that fast or that completely. The rating is the mitigant — but the rating rests on expected parent support, which means a change in SBI's own standing or in how rating agencies view that support would transmit straight into SBI Card's cost of funds.¹¹¹²

Execution and credit-cycle risk has already been demonstrated, which changes how it should be weighted. The 2023–2025 episode showed that real-time delinquency forecasting is a genuine weak point at this company. The mitigating evidence is that remediation worked once identified. But the base rate has been established: when the next unsecured cycle turns, the reasonable prior is that SBI Card will identify it late and fix it competently. A second cycle with the same forecasting lag would be a materially bigger credibility problem than the first, because it would establish a pattern rather than an episode.

Disintermediation risk runs through the RuPay-UPI transition discussed above, and its financial expression is margin rather than volume: the risk is not that spending disappears but that it migrates to rails carrying thinner economics per rupee.

Demand and competitive-substitution risk deserves separate treatment from credit risk, because they are often conflated. The threat to a card issuer is not only that borrowers default; it is that they borrow somewhere else. On the June 2026 call, management identified competition from non-bank lenders in personal loans as a direct constraint on growth in its EMI book — the segment it has been deliberately expanding as the revolver pool shrinks.⁴⁰ That is a specific, current, disclosed competitive pressure on the exact growth vector the company is relying on. A customer who converts a ₹60,000 purchase into a personal loan from a digital lender at a lower rate is a customer whose card sits idle, still counted in cards-in-force, contributing nothing.

Technology and platform risk is real but frequently overstated for this business. Card issuing is not a business where a model breakthrough displaces the incumbent; the barriers are regulatory licences, capital, and collections infrastructure, none of which software erodes quickly. Where technology does bite is in acquisition and underwriting efficiency — a rival that approves a customer in four minutes on a mobile app wins the customer that SBI Card's branch channel would have taken a week to reach. The company's cost-to-income deterioration to 55.3% in FY26 is at least partly the price of keeping pace on digital acquisition and servicing.⁵ The risk here is not disruption; it is a slow, grinding expense race that a monoline has fewer places to absorb.

Cybersecurity, data and IT-governance risk deserves more weight here than at a typical lender, because SBI Card is a payments company that holds card credentials, transaction histories and bureau data for more than 22 million customers. The Kotak Mahindra episode is the sector-relevant precedent: the RBI's remedy for IT risk-management failure was not a fine but a ten-month prohibition on acquiring new customers digitally and issuing new cards.¹⁵¹⁶ For a diversified bank that was painful. For a monoline whose only product is the card, an equivalent order would halt the business. SBI Card discloses company-wide cyber-awareness programmes and carries improving external ESG governance scores, and no such regulatory action against it has been reported in the period reviewed here.⁵ That is not the same as assurance — it is simply the current state of the record.

Accounting judgment worth flagging. Credit card provisioning under Ind AS 109 rests on expected-credit-loss models whose staging definitions and probability-of-default assumptions are management judgments, not observed facts. SBI Card's own disclosure makes the discretion visible: its Stage 2 bucket includes not only accounts 30 to 89 days past due but also accounts flagged as high-risk, over-limit or linked to another delinquent account, and its Stage 3 bucket extends beyond 90-day delinquency to settled, restructured, deceased and linked accounts.⁵ Those are reasonable definitions. They are also elastic ones. The additional provision of ₹220 crore carried above the modelled requirement at March 2026 — up from ₹121 crore at December 2025 — is management exercising exactly that discretion, and in the conservative direction.⁵ An investor should read a rising overlay as prudence and a shrinking one as a question, because an overlay that is released into earnings is a quality-of-profit issue rather than a performance improvement.

Concentration risk of an unusual kind deserves a line. SBI Card's dependence on its parent runs through three separate channels simultaneously: customer sourcing, credit rating, and management. Each is a benefit today. Each is a single point of failure. A change in SBI's strategic posture toward the card business — a decision to push its own in-house card issuance harder, for instance — would affect all three at once. Nothing in the public record suggests this is contemplated, and the promoter's stated position is that the business is core.²¹ But structural dependence of this concentration is a risk factor whether or not it is currently being exercised.

↑ Back to Top

XII. KPIs to Watch & Epilogue

Three numbers carry most of the information about this company, and they should be tracked in order.

First, receivables growth — not spends growth, not cards in force.

This is the KPI that separates the two possible futures. A credit card company that grows spending while its loan book stands still is being paid a toll; a company whose interest-earning receivables compound is being paid a spread. FY26's 2% receivables growth against 29% spends growth is the clearest single indictment in the numbers, and management's own guidance that asset growth picks up from the second half of FY27 is the specific claim to test.⁵¹⁰ If interest-earning receivables — the revolver plus EMI pool, 54% of the book at March 2026 — begin compounding at a rate approaching spends growth, the distribution advantage is converting into value. If they do not, it is not.

One deliberate omission is worth explaining. Cards-in-force market share, the metric that dominates monthly commentary on this sector, is not on this list. It is easy to track, it is published by the RBI every month, and it is close to useless as a measure of value creation — a dormant card counts identically to one carrying a ₹80,000 revolving balance. Spends share is better but, as this cycle demonstrated, can be inflated by low-margin corporate volume. Receivables growth is harder to game.

Second, the gross credit cost trajectory through FY27 and FY28, with particular attention to whether management's guidance holds. The level matters less than the relationship between what was said and what happened. Gross credit cost at 6.5% in the June 2026 quarter is a genuinely good number.⁶ What would make it meaningful is four consecutive quarters in which the outcome lands inside the guided range. That, and only that, would upgrade the assessment of management's forecasting from "poor" to "adequate."

Third, the cost of funds and net interest margin together.

This is the cleanest available read on whether the structural handicap is widening or narrowing. Funding costs at 6.5% and margins at 10.8% in the June 2026 quarter, with the company guiding funding costs higher, describe a margin under pressure from both ends — falling asset yield from the transactor mix shift, rising liability cost from the rate cycle.⁶¹⁰ The comparison worth making is not against SBI Card's own history but against what bank-owned peers report on the same book.

The epilogue is neither triumphant nor damning, and it should not be forced into either shape.

SBI Card enters FY27 as a company that has survived its first serious post-IPO credit cycle without a capital raise, without a governance crisis, and with its market position broadly intact. Asset quality is the best it has been in more than three years. Its balance sheet is over-capitalised rather than stretched. Its acquisition engine has restarted. Its addressable market is genuinely large.

It also enters FY27 as a company with a structurally expensive balance sheet it cannot fix within its current corporate form, an operating cost base growing faster than its income, a lending book that has not grown in a year, a highest-margin customer segment that management expects to shrink slightly as a share of the whole, a leadership model that rotates senior executives on two-year terms drawn from its parent, no outside shareholder with an independent commercial voice, and a regulator that has demonstrated twice in three years its willingness to redraw the economics of this exact business overnight.

Both descriptions are accurate. The IPO of March 2020 priced the first one. The six years since have been the market's ongoing attempt to price the second. What the next two years of receivables growth, credit cost delivery and margin trajectory will determine is which of the two descriptions turns out to have been the more important one.

References

1. SBI Cards IPO details — Chittorgarh ↩↩↩↩

2. SBI Card Shareholding Pattern, 30 June 2025 (PDF) ↩

3. About Us — State Bank of India ↩↩

4. SBI Card Q4 FY25 results — profit slips 19% on higher impairment — Business Standard, 2025-04-24 ↩

5. SBI Card Investor Presentation Q4 FY 2025-26 — Disclosure under Regulation 30, 2026-04-27 (PDF) ↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩↩

6. SBI Card Q1 FY27 slides: credit costs drop, profit jumps 20% — Investing.com, 2026 ↩↩↩↩↩↩↩↩↩↩↩↩↩

7. SBI Cards and Payment Services share price — Tickertape ↩↩

8. State Bank of India and Carlyle To Jointly Acquire GE Capital's Shares in SBI Card — Carlyle, 2017-07-21 ↩↩

9. Carlyle portfolio page — SBI Card investment (status: Exited) ↩↩

10. Earnings call transcript: SBI Cards posts stronger Q1 FY27 profit — Investing.com, 2026 ↩↩↩↩↩↩↩↩↩↩

11. CRISIL Rating Rationale — SBI Cards, March 2025 ↩↩↩

12. ICRA Rating Rationale — SBI Cards (PDF) ↩↩↩

13. India's credit card spends rise 7% to ₹1.97 trillion in April 2026: RBI data — Angel One, 2026 ↩↩↩↩↩↩↩

14. Axis Bank completes acquisition of Citibank India consumer business — Business Standard, 2023-03-01 ↩

15. RBI bars Kotak Mahindra Bank from onboarding customers online, issuing new credit cards — TechCrunch, 2024-04-24 ↩↩

16. RBI lifts business restrictions on Kotak Mahindra Bank — News on AIR, 2025-02-12 ↩↩

17. What is card tokenization — India Briefing, 2022 ↩

18. RBI increases risk weight on unsecured consumer credit exposure — Business Today, 2023-11-16 ↩↩↩

19. RBI Circular — Card Networks Arrangement — Probe42 regulatory updates ↩↩

20. SBI Cards loses credit card market share in July 2024 — Business Standard, 2024-08-30 ↩

21. SBI Card appoints Salila Pande as MD & CEO, effective April 1, 2025 — India Infoline ↩↩↩↩↩

22. About Us — SBI Card leadership ↩↩

23. Appointment of Chief Financial Officer — SBI Card, October 2022 (PDF) ↩

24. The Carlyle Group completes partial exit in SBI Card through IPO — Carlyle ↩

25. Carlyle offloads 4.3% stake in SBI Cards, gets ₹3,944 crore — Business Standard, 2021-03-17 ↩

26. Carlyle to sell 3.4% stake in SBI Cards for $443m — Private Equity Insights ↩

27. Carlyle sells 2.78% stake in SBI Cards for ₹2,229 crore — Business Standard, 2022-04-05 ↩

28. SBI Cards ₹2.50 interim dividend for FY26, record date March 11, 2026 — Angel One ↩↩

29. SBI Card Investor Presentation Q4 FY 2024-25 (PDF) ↩↩↩↩

30. SBI Cards Earnings Call Transcript, Q3 FY 2023-24 (PDF) ↩

31. SBI Cards Q3 FY24 — analysts downgrade on asset quality — Business Standard, 2024-01-29 ↩

32. SBI Cards and Payment Services Limited Q1 FY25 Earnings Call Transcript (PDF) ↩

33. SBI Cards and Payment Services Limited Q2 FY25 Earnings Call Transcript, October 29, 2024 (PDF) ↩↩

34. Brokerages bullish on SBI Card on hopes of credit cost moderation — Business Standard, 2025-01-06 ↩↩

35. SBI Cards and Payment Services Ltd Q4 FY26 earnings call highlights — Yahoo Finance, 2026 ↩

36. How RBI's draft circular on credit risk capital benefits RBL Bank and SBI Cards — Angel One, 2025-10-08 ↩

37. RBI (Commercial Banks – Capital Charge for Credit Risk – Standardised Approach) Directions, 2026 — Taxguru, 2026-04-27 ↩↩

38. 16% of card spends happen on RuPay, half of it on credit via UPI: NPCI — IBEF ↩↩

39. SBI Card second COVID wave — too early to take a call — Business Standard, 2021-06-09 ↩

40. SBI Cards and Payment Services Ltd Q1 FY27 earnings call highlights — GuruFocus, 2026 ↩↩

41. India's RuPay-UPI push is cutting out Visa and Mastercard — TechCrunch, 2025-01-09 ↩