MVB Financial Corp. MVBF

Key data

The business

MVB Financial is a $3.5 billion bank run out of Fairmont, West Virginia, with two businesses stapled together. One is an ordinary community lender: about $1.6 billion of commercial real estate and business loans to customers in West Virginia and Virginia, funded by local deposits. The other is a national deposit machine that banks the companies most banks won't touch. Online gaming operators park cash with MVB, gaming is a named deposit concentration, payment startups and payroll apps route money through it, and it runs an early wage access program plus a book of tax anticipation loans repaid by the IRS rather than by borrowers. In 2025 it sold Victor Technologies, the banking-as-a-service software arm it had built, and booked a large gain. The edge, to the extent there is one, is regulatory: MVB has the compliance plumbing and the appetite to bank fintech and gaming niches, and that brings in low-cost deposits a normal community bank of this size cannot source.

The numbers

The engine is spread income, and it widened. Net interest income grew 25% year over year in Q2 2026 and the tax-equivalent margin reached 4.16% against 3.69% a year earlier. But $2.3 million of that quarter's spread income was non-recurring, and the reported profit leans on gains: the Q3 2025 result came largely from the Victor sale, and Q2 2026 fee income of $18.8 million ran more than double a normal quarter. Strip those out and the earnings power is meaningfully lower than the trailing multiple suggests.

Across five years the record is lumpy, not compounding: earnings per share swing by half from one year to the next with no durable trend, which is what you would expect from a bank leaning on episodic fintech gains. Tangible book value per share, the cleaner scorecard, grew a steadier 12% over the past year to $26.52.

At $31.23 the stock costs about 1.18 times tangible book for a reported return on equity near 12%. That is fair, not cheap, because the 12% is flattered by non-recurring gains in two of the last four quarters. Normalize them and the through-cycle return is closer to 8%, which is roughly what 1.18 times book should pay for. The variant is modest: this is priced about right, and what would settle it is two clean quarters of margin and deposit mix with no one-time help.

Credit is the thing to watch. Noncurrent loans climbed from under 0.6% of the book in mid-2023 to a peak in early 2025, easing since but still elevated, and the provision tells the story plainly: $6.5 million set aside in the first half against $2.2 million a year ago.

The strain sits in the commercial business book, which held $20.7 million of non-accrual loans and $12.7 million more than 90 days past due at quarter-end. Company reserves have built to 1.14% of loans from 0.97% a year ago, roughly keeping pace with the deterioration but no more than that.

Management

Insiders have been buying, in size that is small but one-directional: three open-market purchases worth about $66,000 over the past year, led by $52,000 from director Michael Sumbs last November, and no sales at all. In 2025 the company repurchased $10.2 million of stock, its first buyback in years, done near tangible book, a reasonable price to pay for your own shares. The Victor sale shows a team willing to build a fintech asset and cash it in rather than marry it. The watch-item a recent review flagged, the thin legal-proceedings disclosure, held: neither the August 5 quarterly report nor the August 19 8-K surfaced a new proceeding. The live change instead is the tripled credit provision, which is where attention belongs now.

How it fails or surprises you

Cheap deposits are leaking. Noninterest-bearing deposits fell 6.6% over the year to $1.07 billion while interest-bearing deposits grew 20%, dropping the cheap share from 40% to 34% of the book. The whole bet rests on gaming and fintech money staying put and staying free. If that mix keeps sliding, the 4.16% margin compresses fast once the non-recurring help rolls off. Watch noninterest-bearing share next quarter.

The commercial book could crack. This is the fact the bull read explains least: the provision tripled and non-accruals reached $20.7 million in the business-loan portfolio. If noncurrent loans break back above 1.5% and reserves stop keeping pace, the earnings beat gets eaten by reserve building. Watch noncurrent-to-loans and coverage together.

The fee engine proves durable (right tail). Payment card and service charge income rose 29% to $6.0 million in the quarter, and the efficiency ratio collapsed to 59.5% from 84.7% a year ago. If those payments and gaming fees are recurring rather than one-off, and efficiency holds near 60%, 2027 earnings of $2.51 are beatable and 1.18 times book is too cheap for a low-teens return. Watch the efficiency ratio hold.

Closing thoughts

Two clean quarters will tell you what MVB is. If margin and deposit mix hold without one-time gains, and if noncurrent loans keep falling, then 1.18 times book is cheap for a deposit franchise earning double digits through the cycle. If the margin was propped up by $2.3 million of non-recurring spread income and fee income was a one-off, while noncurrent loans drift higher, then the 12% return is overstated and the price is fair for an 8% business. An ambiguous print, decent headline earnings with more one-time help buried inside, leaves you where you are now, and the honest move then is to keep waiting for the clean quarter. The fatter tail is the downside: the deposit-mix erosion and rising provision are here now and measurable, while the fee-durability upside is still a hope, and if the commercial book turns, the reserve build erases the earnings beat before the upside can arrive.

The bet is still that MVB keeps pulling in cheap deposits from online gambling companies, payment startups and payroll apps, and lends that money out at a wide spread. What breaks it is simple and already moving: the cheap deposits leaking away and the commercial book's noncurrent loans climbing. The one pair of numbers that tells you first is noninterest-bearing deposit share and noncurrent-loans-to-loans. If the first keeps falling and the second keeps rising into next quarter, the wide spread and the clean-credit story are both wrong, and 1.18 times book was not cheap after all.