State Bank of India — the sovereign's bank, finally earning its keep

The Pulse

State Bank of India is the country’s bank — largest, oldest, government-controlled, holding roughly a fifth of all deposits and advances in the system — and after a decade that included real distress, it just printed its best year ever: FY26 net profit of about ₹80,000 crore, return on equity of 18.5%, and bad loans at a two-decade low (gross NPA 1.49%, net 0.39%). The franchise’s defining asset is a deposit base so vast and so cheap (CASA near 40%, ~₹60 lakh crore of deposits) that no one can compete with it on funding cost. The defining constraints are equally clear: SBI earns a thinner return on assets (~1%) than the top private banks because it carries lower-yielding, policy-driven lending and is run partly as an instrument of the state; its capital is leaner than private peers; and its margins are unusually sensitive to RBI’s rate cuts because its loans reprice faster than its deposits. This year it raised ₹25,000 crore in India’s largest-ever share sale — explicitly as “confidence capital,” not because it was short — and is sitting on large, under-recognised subsidiary value (SBI Life, SBI Cards, the AMC). The market pays 1.5 times book for all this, a steep discount to private peers that the chairman himself calls “a conundrum.”

The Business

SBI does everything, for everyone, at a scale that is genuinely hard to picture: over ₹83 lakh crore of group assets, total business above ₹100 lakh crore, more than 50 crore customers, the financial plumbing of the entire country. It earns the usual bank way — spread, fees, treasury — but the engine room is its deposit franchise, which the bank itself describes as “the primary resource-mobilising unit” that funds corporate lending and treasury through internal transfer pricing. That deposit base, gathered through an unrivalled branch and business-correspondent network and the YONO digital platform, is the single most important thing SBI owns: it is the cheapest large pool of funding in Indian banking, and it is effectively un-replicable.

The lending book spans home loans (the retail star, growing ~15% on an ₹8-lakh-crore base), unsecured personal loans (the “Xpress Credit” product, currently a soft spot — flat, as tighter underwriting and a booming gold-loan alternative cannibalise it), SME (strong, ~20%, increasingly underwritten by automated rule engines), agriculture, and corporate. The corporate story this year was deliberately restrained: as rates fell, well-rated companies refinanced bank loans in the commercial-paper market, and SBI chose to let underpriced loans go rather than chase them — while pointing to a sanctioned-but-undrawn corporate pipeline of roughly ₹7–8 lakh crore as the coiled spring.

What makes SBI distinctive cuts both ways. It is a D-SIB — a “domestic systemically important bank,” too big to fail, with the sovereign standing behind it. The Government of India owns about 55.5% (down ~2 points after the QIP). That ownership is simultaneously the ultimate backstop and the ceiling: it means SBI delivers national policy (financial-inclusion schemes, priority lending, green-transition financing tied to “Viksit Bharat 2047”), which is exactly why its return on assets sits structurally below a private bank optimising purely for profit. And folded inside the group are genuinely valuable subsidiaries — SBI Life, SBI Cards, SBI Mutual Fund, SBI General — whose worth barely shows up in the standalone numbers.

How Management Thinks

Chairman C.S. Setty, who took over in August 2024, runs SBI with a degree of candour and through-the-cycle discipline that is unusual for a public-sector bank. The clearest tell is that he keeps beating his own guidance and refusing to raise it. Nine months into a year tracking above target, asked to lift the ROA goal, he declined — “I do not want to jump the gun” — and held the through-cycle ROA target at “above 1%” even as the bank delivered better. He volunteers the unflattering normalisations rather than hiding behind headline beats: when a quarter was flattered by a one-off gain on the bank’s Yes Bank stake, he himself pointed out that ROA excluding it was ~1.04%; he owned the Xpress Credit weakness, the margin dip, and a failed attempt to pass on a deposit-rate cut. For a PSU, that transparency is the franchise’s underrated governance asset.

The strategic mind is organised around two ideas: defend and extend scale leadership while digitising, and manage capital with private-sector rigour. Setty talks constantly about risk-weighted-asset efficiency and a “flywheel” of capital allocation — directing balance sheet to where it earns the best risk-adjusted return, which is why he let underpriced corporate loans walk. On capital itself, the philosophy is revealing: the ₹25,000-crore QIP — the largest in Indian history, oversubscribed several times over — was framed insistently as “confidence capital, not growth capital,” lifting the CET-1 buffer rather than plugging a hole, because SBI was “never capital- or liquidity-constrained.” The preferred route to future capital is not dilution but surfacing subsidiary value — an SBI Mutual Fund (AMC) listing targeted this year, participation in NSE’s eventual listing (an unmarked hidden reserve), and the residual Yes Bank stake. That is a thoughtful, shareholder-aware way to run a state-owned balance sheet.

Credibility is high on delivery — the asset-quality cleanup is real and the numbers back the words — but two honest caveats belong here. First, the government-ownership dimension means capital allocation is never purely a return decision; national-mission lending is part of the mandate, and the owner takes its dividend. Second, the single biggest undisclosed variable is the move to expected-credit-loss (ECL) provisioning, effective April 2027 on a multi-year glide: management has made no provision yet and is withholding the number until after the June quarter, insisting it won’t dent ROA or capital. That promise is the one most worth tracking.

Where It’s Going

The trajectory is steady, high-scale growth with margins as the swing factor. Management guides 13–15% credit growth (retail-and-SME-led, with corporate finally rebounding behind the capex pipeline), a through-cycle ROA above 1%, ROE of at least 15%, cost-to-income below 50%, and credit cost held around 0.5% — a level it has now defended for three years. The margin is the pressure point: as a PSU with a fast-transmitting floating-rate loan book against largely fixed-rate deposits, SBI absorbs rate cuts quickly on the asset side and recovers slowly on the funding side. Full-year FY26 domestic margin was 3.03% — the bank delivered its “above 3%” pledge — but the Q4 exit dipped to 2.93%, and management’s plan to claw it back leans on asset-mix and yield levers (migrating commercial-paper-linked corporate loans back to higher-yielding benchmarks) rather than cheaper deposits.

The genuine tensions: the ECL transition (size unknown), the structural margin sensitivity to the rate cycle, the perennial PSU question of whether thin capital can fund double-digit growth without dilution (subsidiary monetisation is the answer management is betting on), and the soft spot in unsecured retail. None of these threaten the franchise — they shape its returns. The bull case is that a cleaned-up SBI compounding book value at a high-teens ROE, with billions in unlisted subsidiary value, is mispriced. The bear case is that government ownership and the ROA ceiling mean it always will be.

The Four Checks

Quality & moat (gate). Yes — a genuinely good business with a deep, durable moat, with an asterisk. The moat is the deposit franchise: ~22% of all system deposits at a funding cost no competitor can match, reinforced by D-SIB status and the sovereign backstop. That is as wide a moat as exists in Indian banking. The asterisk is that the same ownership that guarantees the moat also caps how much of it converts to shareholder return — SBI is run partly for the nation, not solely for the share register. The moat is real; the question is who captures its economics.
Returns on incremental capital & runway. Respectable and improving, but structurally capped. ROE has reached the high teens (18.5% in FY26, even above 20% in a strong quarter) — but note how: ROA is only ~1%, and the strong ROE is partly leverage on a thin capital base. A private peer earns a similar ROE on a far higher ROA (2%+) and thicker capital. So the quality of SBI’s returns is a notch below, even where the headline matches. The runway, by contrast, is effectively unlimited — SBI grows with India itself.
Capital allocation for the stage. Rational and notably shareholder-aware for a PSU. Letting underpriced corporate loans go, holding credit cost flat through the cycle, raising “confidence capital” pre-emptively rather than in distress, and — crucially — choosing to fund future growth by unlocking subsidiary value (AMC listing, NSE, Yes Bank stake) instead of diluting equity, are all the right calls. The standing constraint is the one no management can allocate away: the government is the controlling owner, sets part of the lending agenda, and draws a dividend. Within that boundary, allocation is being run about as well as a state bank can run it.
Price. Cheap on current returns — demonstrably so — with the discount being the market’s verdict on ownership and the ROA ceiling rather than on the numbers. At roughly 11 times earnings and 1.5 times book for a bank earning a high-teens ROE with two-decade-low NPAs, SBI trades at a steep discount to ICICI (2.5x) and HDFC (~2x), and the chairman openly calls the valuation “a conundrum.” The bull reasoning: the price barely credits the standalone bank, and gives almost nothing for the large unlisted subsidiaries (SBI Life, SBI Cards, SBI MF). The sober counter: PSU banks have traded at this kind of discount for years because government ownership, policy lending, thinner capital, and the looming ECL transition are real, persistent reasons for caution. The price looks undemanding against the fundamentals; whether the discount ever closes depends on factors — ownership, subsidiary value-unlocking — that sit largely outside the bank’s quarterly results.

Sources

Concall transcripts read: Q1 FY26 (Aug 2025), Q2 FY26 (Nov 2025), Q3 FY26 (Feb 2026), Q4 FY26 / full-year (May 2026).
Annual reports read: FY26, FY25, FY24 (trimmed high-signal sections).
Snapshot: screener.in consolidated, fetched 2026-06-09 (logged-out public session).
Note on resolution: the ticker “SBI” initially resolved to SBI Life Insurance in the data source; this digest is State Bank of India (ticker SBIN), re-fetched correctly.
Gaps / caveats: The annual-report trims were heavy on Chairman’s-letter / risk-governance / segment-table content but thin on standalone headline ratios (NIM, GNPA/NNPA, CASA, CAR, government-shareholding line), so the operating figures above are drawn mainly from the four concalls and the snapshot; consolidated vs standalone profit figures differ (group net profit ~₹83,000 cr vs standalone ~₹80,000 cr FY26) and are noted as such. The snapshot did not populate NIM/CASA/CAR fields and carried only stale GNPA/NNPA cells. The April-2027 ECL transition is the key undisclosed forward variable. Full research dumps in vault/Sources/Earnings/State Bank of India/.