ATLC — Q2 2026 Financial Report Analysis

Mercury nearly doubled revenue, but profit grew more slowly than the loan book

Atlanticus Holdings' second quarter of 2026 (three months to June 30, 2026) is the third full quarter that includes Mercury Financial, the near-prime credit card business it bought on September 11, 2025. Total operating revenue rose 89.0% to $744.3 million and diluted earnings per share rose 65.6% to $2.50. Atlanticus doesn't lend to consumers directly. Partner banks (The Bank of Missouri, WebBank, First Bank and Trust) issue the cards and store-financing accounts, and Atlanticus buys the resulting balances and earns the interest and fees on them. The receivables it manages more than doubled, from $3,046.5 million to $6,891.2 million. Of that, $3,054.3 million is the acquired Mercury portfolio.

Profit grew more slowly than the loan book, and the filing explains why. Mercury's cardholders have better credit than Atlanticus' existing customers, so they pay lower interest and default less often. The company's net interest margin fell from 11.7% to 9.3%. Its version of that measure is the yield it earns on receivables minus charge-offs and minus funding costs. A meaningful part of this quarter's profit also came from accounting estimates rather than collections (see below).

At a glance

• Managed receivables: $6.89 billion, up 126%. Excluding Mercury, receivables were $3,836.9 million, about 26% above last year. That is the underlying growth rate of the existing business.
• Net interest margin: 9.3%, down from 11.7%. Atlanticus now earns less per dollar lent after losses and funding. The lower charge-off rate didn't make up for the lower yield on Mercury's lower-risk accounts.
• Diluted EPS: $2.50, up 65.6%. The share count was flat, so buybacks didn't drive this. A $41.4 million gain from revaluing the loan book did a lot of the work.

Key figures

The ratios cover the Credit as a Service (CaaS) segment, which holds almost all of the company's receivables. They are the company's own non-GAAP "managed receivables" measures, based on each loan's face value. The net charge-off ratio is the share of the average loan book written off as uncollectable over a year, after recoveries. Atlanticus calculates its "net interest margin" after deducting charge-offs, which a bank does not do, so it can't be compared with a bank's NIM.

For the first half of 2026, total operating revenue was $1,423.8 million (+92.7%) and diluted EPS was $4.74 (+58.0%).

What drove the quarter

Revenue. Interest and late fees on consumer loans more than doubled to $545.1 million, from $276.4 million. Fee income rose to $150.9 million from $94.3 million. The company attributes $464.3 million of its first-half revenue growth to Mercury. The rest came from its existing programs: active accounts grew by more than 1.0 million year over year, excluding Mercury. Across all programs, serviced accounts rose by more than 2.2 million, including about 1.2 million from Mercury.

Yield fell, by design. The total managed yield ratio (everything earned on the receivables, as an annual percentage of the average balance) fell from 39.1% to 34.2%. Management says Mercury's receivables "tend to have lower yields (and lower associated delinquency and charge off rates)" and are meant to produce "a similarly profitable asset". On this quarter's numbers they haven't yet. Charge-offs fell 2.3 points while the yield fell 4.9 points, so the margin after losses and funding dropped 2.4 points. Management says Mercury's yield should rise as the partner bank rolls out "product, policy and pricing changes", but some of those changes "will take several quarters to be fully realized."

Costs scaled with the book. Total operating expenses rose 91.7% to $157.6 million. The fastest-growing items were marketing and solicitation, up from $24.9 million to $48.1 million, and card and loan servicing, up from $34.1 million to $57.9 million. Interest expense more than doubled to $123.4 million. The company took on Mercury's securitization debt, which was $2,711.4 million of notes payable at June 30, and issued $400 million of 9.75% senior notes due 2030 in August 2025.

The two portfolios moved in opposite directions on credit. In general purpose credit cards, which now include Mercury, the share of balances 90+ days past due fell to 7.2% from 10.0%. Much of that improvement comes from adding Mercury's lower-delinquency accounts to the total, rather than from the same customers paying better. In private label credit (store and healthcare financing), 90+ day delinquencies rose to 4.7% from 3.8%, and from 3.8% in the first quarter. The 10-Q doesn't explain the private label increase.

What the headline numbers hide

• "Provision for credit losses" is not where the losses are. The income statement shows a provision of just $1.0 million. That line covers only the small auto loan book. Most of Atlanticus' loans are carried at fair value, an estimate of what the loans are worth today based on expected future collections. Write-offs on those loans run through a line called "changes in fair value of loans", which was a $396.3 million loss this quarter, up from $216.8 million. Principal and finance charge write-offs alone were $433.3 million, up from $211.8 million.
• Revaluation gains did much of the work. The fair-value line also includes gains when management raises its estimate of what the loans will collect. In Q2 2026 these "favorable changes... in fair value assumptions" added $41.4 million, compared with $17.1 million a year earlier. Separately, a $5.5 million gain came from lowering the estimated contingent payment owed for Mercury (an extra amount payable to the sellers only if the business hits agreed targets). Pre-tax income was $66.0 million. By our arithmetic, excluding those two items leaves about $19 million, below the roughly $23 million calculated the same way for Q2 2025. The gains aren't one-offs in the sense of being fake. The company says they reflect "continued favorable performance of our Mercury portfolio, which was acquired at a lower fair value". But they are estimate changes, and they can reverse if consumer credit weakens.
• The share count didn't help. Diluted shares were 19.17 million, compared with 19.19 million a year earlier. Common stock buybacks were small: 73,004 shares for $3.8 million in the first half. The tax rate was 24.7%, compared with 24.4%. So the EPS growth came from the income statement, mainly the revaluation gains above, rather than from fewer shares or lower tax. Diluted EPS ($2.50) is well below basic EPS ($3.13) because about 4.0 million potential shares from preferred stock exchange rights and stock compensation count in the diluted figure.
• Cash flow isn't a clean check here. Operating cash flow was $563.3 million in the first half, compared with net income of $94.3 million. For a lender like this, operating cash flow mostly reflects finance charges and fees collected. Loan purchases ($3,059.1 million) and principal collections ($2,588.6 million) appear under investing activities, so the gap doesn't show that earnings quality is unusually high.
• Leverage and refinancing. At June 30, the company had $5,578.9 million of notes payable (mostly debt secured by specific loan pools, with no claim on the company's other assets) and $692.1 million of senior notes, against $697.5 million of equity. The 6.125% senior notes mature on November 30, 2026. The company bought back $13.6 million of them in the first half. Management calls its refinancing risk "moderate in the current environment."
• No adjusted EPS. Atlanticus reports GAAP earnings only. Its non-GAAP figures are the receivables ratios in the table, so there's no gap between GAAP and adjusted profit to reconcile.

Takeaway: Atlanticus more than doubled its loan book, but each dollar lent now earns less after losses and funding: the margin fell from 11.7% to 9.3%. The 66% EPS gain leaned on a $41.4 million upward revaluation of its loans. Excluding that gain and the $5.5 million Mercury adjustment, pre-tax profit was lower than a year ago, so the next few quarters need to show Mercury's yield rising as promised.

Outlook

Atlanticus gives no numeric guidance. The 10-Q's MD&A sets out these expectations:

• Growth: general purpose card receivables should keep growing through 2026 and outpace private label. Private label purchases from its largest retail partner should "moderate in the third quarter of 2026", which would make Q3 private label volumes lower than a year earlier.
• Credit: the net charge-off ratio should "continue to marginally improve in 2026". Delinquency rates should rise slightly as the partner banks lend to a broader range of borrowers, though Mercury's lower-risk accounts should limit the increase.
• Margin: the company expects net interest margin to be "consistent... year-over-year." That holds for the half year more than the quarter. The first quarter of 2025 was weak (7.5%), so the first half of 2025 averaged about 9.6%, compared with 9.3% in each quarter of 2026. Q2 alone was 2.4 points lower.
• Costs: quarterly interest expense should keep rising through 2026, and marketing spend should be higher than in 2025.
• Portfolio change after the quarter: an 8-K filed on September 17, 2026 says Atlanticus completed the sale of its CAR auto finance business. That business is small: $77.7 million of auto loans at June 30, out of total assets of $7,491.8 million. The sale price wasn't disclosed in the 8-K.

Our read: the receivables growth is real. Even without Mercury, the book grew about 26%, and the charge-off ratio is lower. The weak point is margin per dollar lent. Mercury's lower-yield, lower-loss cards haven't yet produced the "similarly profitable asset" management describes, and recent profit growth has relied on upward revaluations of the loan book. For the third-quarter report, due around early November, watch three things: whether the 9.3% margin starts rising as Mercury's pricing changes take effect, whether private label delinquencies keep climbing, and whether fair-value assumption gains keep supporting earnings. If the margin stays near 9% while those gains fade, earnings growth will slow to roughly the pace of the loan book or below it.

This is our first published analysis of Atlanticus, so there's no earlier outlook to check this quarter against.