DSIT: Who Brings the Backlog and Who Brings the Profit

The main article already established that underwater is DSIT's engine for revenue and backlog. This continuation isolates the question left open after that first read: if most of the work, most of the projects, and most of the backlog sit in underwater systems, why does most of the segment profit come from RT.

That matters now because 2026 starts with the same basic shape. Backlog at December 31, 2025 was 83.6% underwater and only 16.4% RT, and the company's own 2026 backlog-recognition schedule is still 80.9% underwater. If the gap between the volume engine and the profit engine does not close, more backlog by itself will not fix profit quality.

There is also a short anchor point from the director report. In the second half of 2025, company revenue rose to $22.926 million from $14.467 million in the first half, a 58.5% jump. Operating profit in the same comparison rose only to $1.907 million from $1.596 million, up 19.5%. More volume clearly arrived, but it did not translate into profit at the same pace. Note 24 explains why.

The Real Gap Sits In The Segments

The 2025 numbers are very clear. Underwater generated $26.535 million of revenue, about 2.7 times RT's $9.938 million. But the segment-profit picture was almost the reverse: underwater contributed only $1.427 million, while RT contributed $1.903 million.

That is the heart of the story. Underwater accounted for 79.4% of revenue growth in 2025, yet barely participated in segment-profit growth. Underwater segment profit increased by only $85 thousand versus 2024. RT added $646 thousand. In other words, almost the entire annual improvement in segment profit came from the smaller business line.

And this is not just a one-year quirk. In 2024, underwater segment margin was 9.8% and RT segment margin was 19.4%. In 2025, underwater fell to 5.4% while RT barely moved at 19.1%. The margin gap widened from 9.5 percentage points to 13.8 percentage points.

Why Underwater Brings The Backlog But Not The Margin

The first reason is that underwater growth came through a heavier project layer, not through an engine that converts every extra revenue dollar into profit with the same ease. Note 24 shows underwater revenue almost doubled to $26.535 million from $13.657 million. Section 8.11 adds that $25.702 million, or 68.5% of total company revenue, came from only two product groups: underwater domain awareness systems at $16.534 million and sonar systems for vessels at $9.168 million. The growth engine is therefore very clear, but also very concentrated.

The second reason is that underwater project economics are more exposed to execution variance. Section 8.10.5 lists five very material underwater projects, all multi-year programs with expected completion dates between 2027 and 2032. The most important line in that table is the one that shaped 2025: estimated total costs on project D rose by about $2.5 million as of March 2026, and the company already reflected that impact in the 2025 statements. The result was an 11 percentage-point drop in the project's expected gross margin.

This is not a footnote. It is evidence that in underwater, one fixed-price program can absorb a meaningful part of the year's return. That also fits the broader model the filing describes: long contracts, design and development phases, installation in ports and on ships, acceptance testing, performance guarantees, and multi-year warranty obligations.

Customer structure reinforces that reading. Underwater's two major customers, customer A and customer B, generated $11.688 million and $12.195 million of revenue in 2025. Together that is $23.883 million, or about 90% of segment revenue. The company is careful to say this does not necessarily mean long-term commercial dependence, because these are project customers whose importance changes over time. But at the level of the specific year, 2025 was highly concentrated.

That matters because contract terms are not easy. Customer A sits on agreements from 2022 and 2024 with a combined size of $49.1 million, structured on a fixed-price basis and spanning about 10 years and about 4 years. Customer B sits on two agreements totaling $46.4 million, also fixed-price, with two-to-three-year timelines, delay penalties, and bank guarantees. When that is the engine carrying backlog, high revenue cannot be read automatically as high margin.

The backlog table makes the point even more clearly. At the end of 2025, underwater backlog stood at $83.987 million versus $16.445 million in RT. Expected 2026 revenue recognition from that same backlog stands at $40.333 million in underwater versus $9.508 million in RT.

The implication is straightforward. Even if the company stays busy and revenue recognition remains strong, the 2026 revenue engine still depends mainly on the heavier programs in terms of execution, installation, and milestone risk.

Why RT Delivers More Profit On A Smaller Revenue Base

RT looks different. The segment generated $9.938 million of revenue in 2025, yet held segment profit at $1.903 million and segment margin at 19.1%. While underwater grew much faster and diluted margin, RT grew 53.1% and kept almost exactly the same profitability level it had in 2024.

Several reasons are visible in the filing. First, the company says RT operates mainly as a subcontractor for defense and civilian customers, and that in most cases development and customization are funded by the customer. Second, section 9.10 states that in 2025 RT had no projects meeting the company's "very material" threshold, and no material project groups either. That does not mean there is no risk. It does mean the year was not resting on one single RT program capable of dominating the entire profit picture.

Customer structure in RT is concentrated too, but differently. Israel Aerospace Industries generated $5.091 million of revenue in 2025, and customer D added $3.745 million. Together that is $8.836 million, again about 88.9% of segment revenue. So concentration exists here as well. The difference is that the filing describes a work mix that includes development, repeat production, repairs, and integration into customer systems, rather than one or two large maritime programs dominating the economics.

That difference is visible in the contract model. With IAI, development work is paid by milestones and repeat production is usually paid after delivery and approval. With customer D, the mix includes development projects, repeat production, and repairs, with average engagements of two and a half to three years. That is a different economic profile from multi-year sonar programs and maritime installations.

Still, RT should not be romanticized. Section 9.13.4 states that prices in RT orders are also fixed in advance for the development and production covered by each order, that contracts include liquidated damages for delays, and that development projects sometimes require performance guarantees. RT is not a risk-free segment. What the filing does show is that in 2025 this risk did not translate into a material profitability hit, unlike what happened in underwater.

One more small but important clue comes from section 9.11. Only one RT product group crossed the 10% of company revenue threshold, components for chip-testing systems at $3.745 million. That does not prove full diversification. It does fit the read that RT is generating profit through a narrower revenue base but with less dependence on a single oversized program.

What Has To Happen For 2026 To Look Different

A reader looking only at backlog could reach an overly comfortable conclusion: if RT backlog grew much faster, perhaps it is already on its way to rebalancing the profit pool. That is only a partial read. RT backlog did more than double to $16.445 million from $7.194 million. But underwater backlog is still 5.1 times larger. Even after that growth, 84% of total backlog still sits in underwater.

So 2026 will not rebalance on mix alone. For the profit center to move back toward a more balanced structure, at least one of three things has to happen.

First, underwater programs need to return to a more stable expected profitability profile. After project D, the market will need to see that there is no recurring pattern of estimate resets inside fixed-price work.

Second, new underwater orders need to arrive with better project economics, whether through pricing, milestone structure, or better control over installation and acceptance testing. Without that, more volume will only widen the gap between revenue and profit.

Third, RT has to keep converting development, repeat production, and repairs into revenue without giving up margin. The filing describes a business supported by long customer relationships, integration into existing systems, and follow-on orders. If that holds, RT can keep carrying most of the profit even without becoming the main revenue engine.

There is also a reason to stay cautious. The company states explicitly that backlog scheduling by quarter and year is based on estimates and on milestone completion, which means revenue-recognition timing can shift materially. So even the 80.9% underwater share of scheduled 2026 recognition is not a fixed promise. But until that timing actually moves, it remains the starting point.

Conclusions

The continuation thesis here is very sharp: in 2025, underwater brought most of DSIT's revenue and backlog, but RT brought most of the segment profit.

That does not mean underwater is weak or RT is necessarily the better business. It means the economics of the two segments are different. Underwater runs through larger, longer, and heavier execution programs. RT runs through development, repeat production, and repair work that preserved far better margin discipline in 2025.

The strongest counter-thesis is that the 2025 gap is overstated because it was driven mainly by the hit in project D and therefore does not represent underwater profitability as a whole. That is a serious argument. If project D is a one-off and new backlog arrives with better execution economics, the gap between the two segments could narrow quickly.

But as of year-end 2025, that remains only a counter-thesis. The reported numbers still say underwater provides the scale, RT provides the profit line, and 2026 begins with a backlog mix that does not solve that by itself. The central question in the next reports will therefore not be whether there is enough work. It will be where the margin actually stays.