BankUnited, Inc. (NYSE: BKU) used its second-quarter print on July 22, 2026 to mark an inflection that management has been telegraphing for several quarters. The Miami Lakes-based regional bank reported net income of $70.7 million, or $0.97 in diluted earnings per share (EPS), up from $61.9 million and $0.83 in the first quarter of 2026 and from $68.8 million and $0.91 a year earlier. Pre-provision net revenue (PPNR, a non-GAAP measure that captures revenue minus operating expenses before setting aside loan-loss reserves) of $109.9 million was essentially flat sequentially and versus a year ago, but the bottom line moved because the provision for credit losses fell to $15.6 million from $24.6 million in the first quarter. Return on assets (ROA, net income divided by total assets) improved to 0.81% on an annualized basis from 0.72%, and return on equity (ROE, net income divided by shareholder equity) climbed to 9.3% from 8.1%, with the leverage reflecting ongoing share repurchases and book-value compounding.
The story underneath the numbers is a deliberate balance-sheet reshape. BankUnited cut its wholesale funding by $1.4 billion versus the first quarter and versus a year ago, and grew non-interest-bearing demand deposits (NIDDA) to a record $9.9 billion at quarter end, or 34.4% of total deposits, the highest mix in the company's history. Net interest margin (NIM, which is net interest income divided by average earning assets) expanded 7 basis points sequentially to 3.06% and 13 basis points from a year ago. We read this as a franchise with mid-single-digit balance-sheet momentum being repriced into a cost-of-funds advantage that should persist while the Federal Reserve remains on hold or cuts further.
Our interpretation is that BKU has quietly exited the high-cost, brokered-funded growth model it ran in 2022 to 2023 and is now compounding fee-rich, deposit-funded earnings. The market appears to be pricing BankUnited as a slow-growth Florida bank trading around book value, but the combination of a 3.06% NIM, an allowance for credit losses (ACL) to non-performing loans (NPLs) coverage ratio of 97.14%, and an $50.1 million quarterly buyback at an average price well below tangible book value suggests the franchise is generating more intrinsic value per quarter than the share price reflects. The single load-bearing risk is a Florida commercial real estate (CRE) office and land book that remains the largest source of criticized and classified loans; the falsifiable clock is the third-quarter 2026 earnings release, where any reacceleration in net charge-offs above the current 0.11% annualized rate would test the credit-quality thesis.