Customers Bancorp, Inc. CUBI

Key data

The business

Customers Bancorp is the holding company for Customers Bank, a roughly $28-billion-asset commercial bank run out of suburban Philadelphia. It earns money the plain way: gather deposits from businesses, lend to other businesses, keep the spread. What sets it apart is how it grew. When Signature Bank failed in March 2023, Customers hired about ten of its venture and fund-banking teams, and with them came low-cost commercial deposits from private-equity and venture funds. That deposit engine, alongside specialty lending in equipment finance, mortgage warehouse (short-term credit to mortgage lenders), and fund finance, is the franchise. It wound down its old crypto digital-asset deposit business, which once funded a large slice of the balance sheet and drew a Federal Reserve enforcement action in 2024. The moat, such as it is, lives in those banking teams and the relationships they carry: a switching-cost franchise only as sticky as the bankers who hold it.

The numbers

Earnings have run well above the reported fiscal-year pace since a weak first quarter of 2025 dropped out of the count. The five most recent quarters:

Net interest income has held near $200 million a quarter while the cost of funds fell to 2.64%, a favorable deposit-mix shift the bank called out directly. Quarterly EPS beat consensus in all four prints. Over five fiscal years the record is lumpier:

The lens table is credit, and it is the one number moving the wrong way:

Read plainly: non-accruals rose 28% in six months and non-performing assets climbed 18%, driven by multifamily loans, where the bank built reserves against higher charge-offs, and by consumer installment loans. The absolute level is still low for a bank this size, and the half-year provision of $46.4 million actually fell from $49.1 million a year ago. The tell is direction, not altitude: the newest, least-seasoned books are the ones starting to leak.

On compounding, equity grew from $1.84 billion at the end of 2024 to $2.21 billion at mid-2026 while the share count fell from about 35.6 million to 33.8 million on buybacks. Book value per share is approximately $65, so the stock at $81 pays about 1.2 times book value. Trailing earnings run near $8.20 a share, well above 2025's $6.26 because the year-ago first quarter carried one-time charges that have since rolled out of the count. Normalize that out and this is a mid-teens return on tangible common equity. Consensus sees $8.45 this year and $9.68 next. The variant is narrow: the market prices the hired-team deposits as rented rather than owned, and the print that settles it is whether funding cost keeps falling as deposits grow.

Management

Customers is founder-run: Jay Sidhu built the bank and chairs it, his son Sam Sidhu is CEO. That family alignment cuts both ways here. Over the last year insiders sold $17.2 million of stock and bought $73 thousand, with Jay Sidhu alone selling about $10.1 million across May and August 2026, plan status not disclosed. Against that, the company repurchased $42.7 million of stock in the first quarter of 2026 under a fresh authorization, after buying back only $5.6 million in all of 2025, and did it near 1.2 times book, which is accretive. The founder selling into strength while the company buys shares is the tension to sit with. The 2024 Fed consent order over the crypto-deposit unit is the scar that argues for watching this team's risk controls, not just its growth.

How it fails or surprises you

Credit turns from creep to cycle. The multifamily and consumer installment books are where charge-offs are building, non-accruals moved from 0.27% to 0.34% of loans in six months, and non-performing assets rose to $85.7 million. If those books accelerate into a softer economy, provisions swamp the earnings growth. The October quarter's charge-off line is the first honest read, and nothing filed since early September has moved it.

The teams were renters, not owners. The deposit franchise rests on bankers hired from Signature. Deposits grew to $21.7 billion and cost of funds fell to 2.64%, which argues the relationships are sticking. But a jump in deposit cost or an outflow quarter would reveal the teams could walk, and take the funding with them. Cost of funds against deposit growth is the gauge.

The deposit machine re-rates (right tail). If funding cost keeps falling as the venture-banking mix matures and credit stays contained, net interest income compounds and a bank earning a mid-teens return on tangible equity does not stay at 1.2 times book. Consensus of $9.68 next year could prove low. The trigger is one quarter of falling funding cost with flat charge-offs.

Closing thoughts

One print settles it: whether the credit creep accelerates or holds. At about 9 times earnings and 1.2 times book, the stock already discounts the rented-team risk and a normal credit cycle. What it does not price is which way the multifamily and consumer charge-offs break. Hold near current levels and the mid-teens return re-rates higher, accelerate and the earnings growth is already spoken for. The left tail is a credit cycle landing on the newest loans just as the founder trims his stake. The right tail is a durable low-cost deposit machine the market still treats as borrowed. On balance the fatter tail is up, because funding cost is falling and the balance sheet is better capitalized than it was, but the credit creep is real and unresolved.

The bet is still that Customers Bank gathers business deposits cheaply and lends them profitably, growing net worth per share at low-to-mid teens while the stock sits near that net worth. What breaks it is the loans, not the deposits: watch net charge-offs in multifamily and consumer installment against the cost of funds. If charge-offs accelerate while funding cost stops falling, the compounding stalls. If funding cost keeps dropping and bad loans hold near 0.34%, you are paying nine times earnings for a mid-teens compounder, and the October charge-off line is the next test.