Matrix: How much of 2025 cash generation is actually repeatable

The main article already argued that Matrix's most sensitive number in 2025 was not revenue and not even margin. It was cash quality. This follow-up isolates only that question: how much of the NIS 835.3 million of operating cash flow in 2025 came from the business's recurring cash engine, and how much was pulled forward through a year-end receivables assignment.

The answer is sharper than the headline suggests. Matrix's cash balance is real, but 2025 cash flow was less clean than the headline number. The fourth-quarter receivables assignment brought in about NIS 261.4 million, a sum larger than the entire year-over-year increase in operating cash flow. Anyone reading 2025 as proof that Matrix now runs at an annual operating-cash level above NIS 800 million is reading the report too aggressively.

Four points frame the issue:

• The full year-over-year increase in cash flow is explained, and more than explained, by one year-end move. Operating cash flow rose to NIS 835.3 million from NIS 619.2 million, an increase of NIS 216.1 million. The receivables assignment alone brought in about NIS 261.4 million.
• The receivables cleanup was only partial. Trade receivables and accrued income, net, fell by NIS 127.3 million, but within that line billed receivables fell by NIS 164.3 million while accrued income rose by NIS 38.1 million to NIS 512.8 million.
• After real cash uses, far less remained free. After fixed-asset spending, acquisitions, lease principal, debt amortization, and cash dividends to shareholders, 2025 left roughly NIS 144.7 million. Strip out the receivables assignment and the picture flips to a roughly NIS 116.7 million deficit.
• The trust deed is barely the constraint. Series 2 blocks distributions only if net financial debt to balance sheet rises above 37%, net financial debt to adjusted EBITDA rises above 4, or equity falls below NIS 450 million. At year-end 2025 Matrix was sitting on net cash of NIS 149.9 million and equity of NIS 1.215 billion.

Most Of The Annual Cash-Flow Lift Was Concentrated In Q4

The first number that matters is not just the annual cash-flow total, but how concentrated it was. In the fourth quarter alone, Matrix generated NIS 566.5 million of operating cash flow, versus NIS 335.0 million in the prior-year quarter. That is almost 68% of the full year's operating cash flow. In that same window, the company also entered into a receivables-assignment transaction with a financial institution for about NIS 261.4 million, structured as a true sale, and derecognized the financial asset from its books.

That is the core of the story. The assignment by itself is larger than the full annual increase in operating cash flow. So even without assuming anything beyond what the filings say, one conclusion is straightforward: the reported 2025 number does not by itself represent the recurring cash run-rate of the business.

The third bar is not a reported number. It is a simple analytical bridge: reported operating cash flow less the consideration received for the assignment. It is not meant to claim that NIS 573.9 million is a formal adjusted metric. It is meant to show how differently the year reads once the year-end transaction is removed from the picture.

Put differently, 2025 was still a good cash-flow year. But it did not prove that Matrix is already producing more than NIS 800 million of recurring operating cash on its own. It proved that the company was able to end the year with a very strong cash-flow number, partly by monetizing receivables.

The Receivables Line Improved, But Not Across The Whole Stack

If one looks only at the balance-sheet headline, the story appears simple. Trade receivables and accrued income, net, fell to NIS 1.799 billion from NIS 1.926 billion. But the internal composition of that line tells a more precise story.

Billed receivables, including shekel, foreign-currency, related-party balances, and collection drafts, fell by a combined NIS 164.3 million. Accrued income, however, rose to NIS 512.8 million from NIS 474.6 million. In other words, the improvement came mainly from the receivables layer that had already been billed and could be collected or assigned faster, while the unbilled revenue layer did not shrink. It grew.

This does not mean the assignment was somehow illegitimate. It is a disclosed transaction, and the company explicitly presents it as a true sale. But it does mean the balance sheet did not suddenly become light on working capital. Part of the cleanup came from the part of the receivables book that was easiest to turn into cash, while the unbilled layer still moved higher.

The average credit figures make the same point in a broader way. In the directors' report, Matrix shows average customer credit of NIS 1.908 billion against average supplier credit of NIS 888.3 million. That is a gap of a little over NIS 1.0 billion that the business still has to finance through operating cash, equity, and credit lines. A year-end transaction can improve the closing snapshot. It does not by itself change the economics of the cycle.

That is exactly what a first-read reader can miss. A decline in the receivables line does not automatically mean structural improvement in cash conversion. Sometimes it simply means cash was accelerated out of an existing receivables book. Here, the split between billed receivables and accrued income supports that interpretation.

Two Cash Frames, And Only One Of Them Is Truly Repeatable

To read 2025 correctly, two different cash frames need to be kept separate:

That bridge is intentionally tougher than the headline. It does not try to estimate maintenance capex, it does not rewrite the financial statements, and it does not invent a management-style adjusted cash metric. It simply asks how much cash remained after the uses that actually happened.

That is also where the distribution question enters. On a cash-flow basis, NIS 269.1 million went out in 2025 as dividends to shareholders. At the same time, the investor presentation already shows another NIS 73.1 million dividend for Q4 2025, or NIS 0.79 per share, equal to about 75% of the combined quarterly profit of Matrix and Magic. So this is not a theoretical debate. Management is already treating this cash profile as a distribution base.

That does not automatically make the policy aggressive. Matrix ended 2025 with NIS 902.9 million of cash and equivalents, NIS 1.235 billion of undrawn facilities, and NIS 2.138 billion of total liquid sources. The issue is not liquidity pressure. The issue is interpretation. If 2025 is read as a year of NIS 835 million of recurring operating cash, a 75% payout looks close to automatic. If it is read as a year in which part of the cash was pulled forward through receivables, the distribution still looks comfortable, but it rests more on balance-sheet flexibility and less on self-repeating cash generation.

The Trust Deed Is Not The Brake, Which Makes This A Management Question

This is where legal restriction and economic restriction need to be separated. The Series 2 trust deed sets a relatively permissive distribution framework. A distribution is blocked only if net financial debt to total balance sheet rises above 37%, if net financial debt to adjusted EBITDA rises above 4, if post-distribution equity falls below NIS 450 million, if the company is in breach of another material undertaking to bondholders, or if an immediate-repayment event exists. The deed also states that, beyond those limits and subject to law, no additional restriction on distributions applies.

So the trust deed does not really explain the debate. It does not corner Matrix. If anything, it leaves the company with wide flexibility. The February 2026 Series 2 issuance reinforces that reading: around NIS 300 million, five-year maturity, 0.5% interest, with proceeds mainly intended to repay more expensive debt at Magic. This reads more like a balance-sheet optimization move than one aimed at solving immediate distribution pressure.

And that is exactly why cash quality becomes the main question. When covenants are not tight, management becomes the real brake. The decision on how far to keep distributing will not be dictated by the trust deed. It will be dictated by how the board chooses to interpret 2025 cash: as a recurring base, or as a year in which part of the cash was accelerated out of the receivables book.

Conclusion

Matrix did not "buy" 2025 cash at a dangerous price. That is not the thesis. The company enters 2026 with a strong balance sheet, net cash, large facilities, and a cheap new bond. But there is a major difference between saying the company is liquid, and saying that 2025 proved a new, sustainable distribution base.

The more conservative reading still looks right. The recurring cash engine of the business was probably materially lower than NIS 835.3 million, and the amount left after real cash uses was much smaller than the headline implies. That is what makes 2026 the real test. If the next reports show that Matrix can keep converting profit into cash without another large receivables move, 2025 can be read as the start of structural improvement. If not, 2025 will look more like a year of sophisticated working-capital management and less like a lasting step-up in cash generation.