Glacier Bancorp, Inc. GBCI

Key data

The business

Glacier Bancorp runs banking the old way: not one bank but roughly sixteen separately branded divisions, First Security in Bozeman, Bank of Idaho, Mountain West, each keeping its local name, lenders and board while Glacier centralizes the back office, technology and credit rules. About 230 branches across eight Mountain West states. The product is a plain community-bank balance sheet: some $21.4B of loans to local businesses, ranchers, builders and homeowners, backed by $6.5B of securities and a deposit base gathered branch by branch. The moat is that deposit franchise, small sticky low-cost accounts in towns where Glacier's division is the bank families have used for a generation, which funds loans cheaper than a national rival can. Growth comes two ways: lend a little more each year, and buy a small bank every year or two, keeping its name and its customers. The detail a customer would recognize is the sign over the door, which still reads First Security, not Glacier. That deliberate localness is the strategy.

The numbers

The story in the numbers is a profit recovery. EPS fell from $2.86 in 2021 to a $1.68 trough in 2024 as funding costs outran loan yields, then turned.

Every line is inflecting up. Net interest income, what the bank keeps after paying depositors, rose 33% from a year ago to $276M last quarter, and net income jumped 85% to $98M, the biggest quarter Glacier has posted. Diluted EPS of $0.75 edged consensus, and first-half EPS of $1.38 rose 48%. The driver is a widening margin as deposits repriced down faster than loans, the reverse of 2023 and 2024.

Run it forward. Last quarter annualizes near $3.00 a share, and consensus sees $3.04 in 2026 and $3.61 in 2027. Return on tangible common equity is back to about 14%, and tangible book per share compounded to $21.81, up 10% in a year even after Glacier issued $605M of stock late in 2025 to fund a deal. At $46.83 the shares change hands near 2.1 times tangible book, the low end of a twelve-year range that ran from 2.1 to 3.0 and usually sat between 2.3 and 2.8. A below-average price on recovering returns.

The caution sits in credit.

Reserves have held flat at 1.22% of loans for six straight quarters while the trouble underneath built: non-performing assets rose from 0.17% of subsidiary assets to 0.29% in a year, and noncurrent loans at the bank charter climbed from 0.20% of loans in mid-2023 to 0.37% by early 2026. Charge-offs stayed near zero, so this is early seasoning, not loss. Yet Glacier cut its first-half provision to $12.4M from $28.1M a year earlier as those late loans grew. That is the number the recovery explains least well: the falling provision that flatters this year's EPS is what a bank sets only when it believes the credit creep is noise. What this memo believes the market does not is that the margin recovery holds and the creep stays benign, making $3-plus earnings at a low-end tangible-book multiple the mispricing. The single print that settles it is the third-quarter provision against the non-performing trend.

Management

Chesler and his team run Glacier as capital allocators first: a decade of steady tangible-book-per-share growth funded by disciplined, stock-paid purchases of small franchises, and the late-August 2026 deal announcement shows the engine still running. Insiders back it with cash, nine open-market buys worth $479,000 over the past year and not one sale, including CEO Randall Chesler's near-$100,000 purchase last November and director Craig Langel's $106,000 buy on August 28, the most recent Form 4 on file. That buying, the item flagged a week ago, held: one-directional and intact. The open question is the acquisition math, whether each deal is bought below the tangible book it adds.

How it fails or surprises you

Credit creep turns into a cycle. Noncurrent loans rose from 0.20% to 0.37% of loans over three years and non-performing assets to 0.29% of assets, yet Glacier cut its first-half provision to $12.4M from $28.1M and held reserves flat at 1.22%. If this is the front edge of a downturn, reserves rebuild and the earnings recovery reverses. The tell is the third-quarter provision.

The margin recovery stalls. The entire profit jump is a widening spread, net interest income up 33% as deposit costs fell faster than loan yields. Deposit interest expense already flattened near $72M a quarter. If competition reprices funding back up, or the $3.0B held-to-maturity book locks in low yields, the spread narrows and EPS drifts toward its 2024 trough.

Rate cuts and the roll-up compound (right tail). If the Fed eases while acquired books season, deposit costs fall further, the margin widens again, and Glacier keeps buying small banks below the tangible book they add. Consensus $3.61 for 2027 would prove light, tangible book would compound double digits, and a stock at the low end of its multiple re-rates toward its usual 2.5x. Watch the next deal's price.

Closing thoughts

The third quarter's credit report settles it: the provision against the non-performing trend. The market pays a below-average 2.1 times tangible book because it doubts the recovery is clean, suspecting the low provision borrows from tomorrow. On the other side of that doubt is a bank earning about 14% on tangible equity with a cheap, sticky funding base and a decade of compounding book value. An ambiguous print, non-performers up but charge-offs still near zero, leaves you holding a fairly priced compounder. A clean one, the provision still low while non-performers flatten or fall, turns the low multiple into the mispricing. The fatter tail is up, because the downside here is a drift back to fair value, not permanent loss, while the upside is a re-rate on a franchise that has survived every Mountain West cycle since 1955.

The bet is still that Glacier's small-town banks lend for more than they pay depositors and keep buying little banks to grow, and that this year's profit is earned, not borrowed from under-reserving. What breaks it is the pair to watch each quarter: the provision against the non-performing trend. If non-performers keep climbing while the provision stays cut, the recovery was an accounting mirage. If charge-offs hold near zero as the margin stays wide, $3-plus earnings are real, and the cheapest tangible-book multiple in twelve years will not last.