Peninsula in the First Quarter: Profit Held, but Project Finance Still Absorbs Provisions

Peninsula did not report a bad quarter, but it also did not give the market the clean proof it was looking for after the 2025 annual results. Net profit barely moved, funding costs declined, and the credit portfolio had already grown again near the approval date of the financial statements, but credit quality has not yet clearly turned. The good news is that the problematic Tomer Levy loan was fully repaid after the balance-sheet date, closing in cash one of the main risks from the prior coverage. The less comfortable part is that project finance, which the company presents as a growth engine, produced higher revenue but almost no operating profit because of higher provisions and a new defaulted project of about NIS 21 million. At the same time, Meitav Investment House increased its holding to 87.83% through the exchange tender offer, but did not complete a delisting, so the discussion shifts back from corporate structure to credit quality and dividends. All-in cash flexibility after actual cash uses was narrow: operating cash flow of NIS 10.8 million almost exactly covered the dividend, lease repayment and small investments, while the next dividend has already been approved. The coming quarters will show whether the old-file repayment marks the start of a broader normalization, or only closes one issue while project finance keeps consuming a large part of the profit it creates.

Company Overview

Peninsula is a non-bank lender to Israeli businesses. Its economics are not a plain growth story. They combine three engines: financing spread on the credit portfolio, control of credit losses, and recurring cash returns to shareholders. In this kind of company, stable net profit can look good, but earnings quality is mainly determined by who pays for it: customers that pay on time, projects that advance without default, or the balance sheet absorbing provisions.

In the first quarter, the core activity remained business credit, with NIS 1.49 billion of net credit. Alongside it stood construction project finance, with NIS 226.9 million of net credit, and financial guarantees with NIS 298 million of off-balance-sheet exposure. That map matters because the company is not only increasing its credit portfolio. It is gradually changing the type of risk it takes.

What differentiates Peninsula from a routine lending company is not leverage or reliance on credit lines. Those are normal in the sector. The unusual point is the gradual shift toward real-estate-backed credit and project finance, while the new engine's profitability still does not match the risk weight it is receiving. That was the yellow flag in the March analysis of credit risk in the portfolio, and the first quarter did not close it.

Provisions Fell, but Credit Quality Did Not Fully Reset

The simple headline for the quarter is net profit of NIS 17.1 million, compared with NIS 17.5 million in the parallel quarter. Financing income declined from NIS 54.4 million to NIS 51.3 million, and net financing income declined from NIS 38.1 million to NIS 36.4 million. Lower financing expenses, down from NIS 16.3 million to NIS 15.0 million, helped hold profit, and credit-loss expenses edged down from NIS 5.2 million to NIS 4.8 million.

The more important comparison is with the fourth quarter of 2025, when credit-loss expenses jumped to NIS 11.6 million. Against that pressure point, the first quarter looks like a return to a more reasonable level. Still, it is not a full reset, because the composition changed: in business credit, provisions fell from NIS 5.1 million in the parallel quarter to NIS 2.9 million, but in project finance they rose from NIS 0.1 million to NIS 1.9 million.

The business-credit note gives a two-sided picture. On the positive side, the Tomer Levy loan, with a balance of about NIS 39.6 million and secured among other things by a first-rank mortgage over real-estate assets, was fully repaid on May 19, 2026. That effectively closes an exposure that weighed on the year-end 2025 read, and weakens the counter-thesis that the specific file would remain open for a long period.

On the other side, at the balance-sheet date, impaired business-credit debt was still NIS 212.5 million gross, nearly unchanged from NIS 214.2 million at the end of 2025. The total allowance in business credit rose to NIS 86.3 million, from NIS 83.3 million at year-end, and the total expected-loss rate increased from 5.05% to 5.48%. The arrears composition is also not clean: total overdue, rescheduled and past-due balances declined to NIS 142.7 million, but balances overdue by 91 to 180 days rose to NIS 23.3 million, from NIS 7.2 million at year-end 2025.

The conclusion needs to be precise. The first quarter closed one known risk, and did so in the best possible way for a lender: actual collection. It still did not prove that the overall cost of risk has returned to a low and durable level.

Project Finance Is Generating Revenue, but Provisions Absorb the Profit

Project finance is the reason this quarter cannot be read as a clean recovery quarter. Segment revenue rose from NIS 4.3 million in the parallel quarter to NIS 5.9 million, and net financing income rose from NIS 3.0 million to NIS 3.9 million. Those numbers show demand and a real growth engine.

The problem is that almost all of the improvement disappeared before reaching profit. Credit-loss expenses in the segment rose to NIS 1.9 million, and operating profit fell to only NIS 0.3 million, compared with NIS 1.1 million in the parallel quarter. In business terms, the segment is selling more credit, but too much of the income is already being absorbed by risk recognition.

The key number in the table is not only the rise in defaulted loans, but the pace of change relative to portfolio size. The project-finance book barely grew versus year-end 2025, but defaulted loans increased by about NIS 20.2 million, and the allowance rose by about NIS 1.9 million. The immediate reason is one project classified during the quarter as defaulted, with a balance of about NIS 21 million, for which the company recorded a provision of about NIS 1 million.

That is an unusual point for the non-bank lending model because project finance should not be measured only by portfolio growth. In this activity, collateral, sales rates, project surplus and monthly monitoring are supposed to provide a protective layer. If the expected-loss rate still rises this quickly, the market should not ask only whether Peninsula can grow the segment. It should ask whether the next phase of growth comes in better projects, with more dispersion and a profit rate that compensates for the risk.

Conclusions

Peninsula exits the first quarter with a mixed but not weak read. Profit held, the problematic Tomer Levy loan was fully repaid, and the funding structure is far from covenant pressure. The blocker is that improvement has not yet come from the most important place: a broad and consistent decline in credit risk, especially proof that project finance can grow without consuming its profit through provisions.

The dividend sharpens that point. The company paid NIS 9.8 million in March and approved another dividend of about NIS 10.5 million after the balance-sheet date. First-quarter operating cash flow was NIS 10.8 million, and after NIS 0.2 million of investments, NIS 0.6 million of lease principal repayment and the dividend paid, cash increased by only NIS 0.2 million. This is all-in cash flexibility after actual cash uses, not normalized cash generation, and it shows that the payout depends on credit losses not returning to the fourth-quarter level.

The tender offer also did not remove credit quality from the agenda. Meitav Investment House increased its holding to 87.83%, but the forced-sale condition was not met and Peninsula was not delisted. The April analysis of the tender offer treated the acceptance rate as the line between a structural event and normal public-company monitoring of a lender. That event was not completed, so shareholders and bondholders return to the same test: whether the portfolio is actually improving.

The next quarters need to show three signs: lower impaired debt after the repaid loan is reflected, no further growth in defaulted project-finance loans, and more comfortable dividend coverage from cash flow that does not rely only on credit-book or credit-line movements. If those arrive together, 2026 can look like a normalization year after a difficult fourth quarter. If not, stable profit will matter less than how much problematic credit still has to pass through the income statement.