ImageSat: How Much of 2026 Backlog Is Truly Locked In, and How Much Can Still Slip
The main article already argued that insurance stabilized the balance sheet. This follow-up shows that 2026 does have a real contractual base of $78.4 million from backlog, but the quality of that lock-in is uneven: Customer A and Chile together account for 71.6% of backlog, while the project-heavy layer is still exposed to milestones and timing.
The main article already framed ImageSat's turning point: insurance proceeds bought the company time, and the debate now shifts from balance-sheet survival to backlog conversion. This follow-up isolates a narrower and more important 2026 question. Not whether demand exists, because it clearly does, but how much of next year is backed by contracts that give the company a real floor, and how much can still move to the right because of timing, milestones, and customer concentration.
At first glance, the setup looks strong. Year-end backlog stood at $160.8 million, and the company maps $78.4 million of that into expected 2026 revenue. The investor presentation goes further and says that backlog alone implies roughly 28% year-on-year revenue growth versus 2025. But the same evidence set also says that some project milestones originally planned for the second half of 2025 will only be completed and recognized in the first half of 2026. That is the right starting frame: the backlog is real, but 2026 is not "locked" in the simple headline sense.
What Is Actually Locked In For 2026
The strongest positive point is that the company is not talking here about pipeline or commercial discussions. It defines backlog as binding orders that have not yet been recognized as revenue. So the $78.4 million mapped into 2026 is not a marketing aspiration. It is a revenue-recognition schedule built from signed backlog. More than that, the same timing table is presented both as of December 31, 2025 and close to the filing date, March 30, 2026, without change. In other words, by the end of March the company had not disclosed any new revision to that recognition profile.
The 2026 mix is also better than the headline fear suggests. Of the $78.4 million scheduled for 2026, about $37.4 million comes from satellite services and another $18.3 million from intelligence as a service. That means roughly 71% of the 2026 recognition plan comes from the company's more service-oriented lines, while only about $22.6 million, roughly 29%, sits in satellite solutions and supporting infrastructure. The long-dated tail looks different. In 2028 and beyond, $13.7 million out of $22.8 million, about 60%, already sits in the more project-heavy line. That matters because it means the near-term year is less project-heavy than the outer years.
Customer disclosures point in the same direction. Customer A carries an analytics agreement worth about $54.5 million, and in 2025 the company also signed a two-year EROS constellation services, equipment, and support agreement worth a total of $42 million. Customer B operates under a five-year agreement worth about $37.5 million, with an option to extend for another five years. None of that removes risk, but it does explain why not all of the 2026 backlog looks like one-off project revenue.
Where 2026 Is Still Exposed To Slippage
The problem starts with concentration. The year-end backlog split is $59.3 million for Customer A, $55.8 million for Chile, $21.2 million for Customer B, $14.8 million for Customer F, and $9.7 million for others. Customer A and Chile together account for $115.1 million, or about 71.6% of total backlog. The same dependence already shows up in 2025 sales: 48% of revenue came from Customer A and 22% from Chile. So the core issue is not only how much backlog exists, but how much of the year effectively sits on two names.
The company explicitly acknowledges dependence on Customer A. Its mitigation argument is that if Customer A were to exit, the company could sell the same area-of-interest capacity to other customers because demand there is high. That may be a reasonable market argument, but it is not the same thing as contractual diversification inside the current year. Replacing one customer with another may be possible over time and still leave 2026 exposed if recognition timing or service volumes shift unexpectedly.
Chile is even more clearly timing-sensitive because the contract has already been re-timed once. The Chile national space program was originally signed at about $109.9 million over roughly five years, with an option to expand by another $9.5 million. In the fourth quarter of 2024, the customer notified the company that it was canceling the scope related to sales of service from the first RUNNER satellite, worth about $3.9 million, due to delays in delivery. Then, in the contract update signed on January 30, 2025, the parties updated the work plan and extended the project through 2028, updated the payment schedule, and accelerated the reduction of advance and performance guarantees.
That update cuts both ways. On one side, it improves the financing and execution profile because guarantee relief is accelerated. On the other side, the need to extend the work plan through 2028, after a canceled scope tied to delays, is hard proof that signed Chile backlog should not be treated as if timing were mechanically fixed. The contract is real. The timing has already proved movable.
| Backlog Layer | What Supports Lock-In | What Keeps It Exposed |
|---|---|---|
| Customer A | Multi-year agreements, long operating history, no collection issues over the years | Explicit dependence on one customer, 48% of 2025 sales and $59.3 million of backlog |
| Chile | Large signed contract, updated payment schedule, faster guarantee step-down | Work plan extended to 2028, $3.9 million scope canceled because of delays |
| The rest of 2026 | Fairly even quarterly schedule of $18.7m, $22.3m, $19.1m, and $18.2m | The company already said some 2H 2025 milestones slipped into 1H 2026 |
The quarterly schedule for 2026 looks balanced, $18.7 million in the first quarter, $22.3 million in the second, $19.1 million in the third, and $18.2 million in the fourth. That is a positive. But this is exactly where the yellow flag sits. The first half of 2026 is also supposed to absorb milestones that slipped out of the second half of 2025. So the first real test of backlog quality will not be the year-end number. It will be first-half recognition.
Why The Pipeline Decline Matters For Backlog Quality
Order pipeline is not backlog, and the two should not be blurred. But pipeline is still the support layer behind the current story. Here the picture weakened. The company shows a pipeline of about $1.0 billion versus about $1.55 billion a year earlier, a decline of roughly 35.5%. At the same time, about $0.82 billion, around 82% of that total, comes from existing customers, while only about $0.18 billion comes from new ones.
That does not mean demand disappeared. A $1.0 billion pipeline is still very large relative to a $160.8 million backlog. The point is different. The cushion behind the current backlog is smaller than it was a year ago, and it is still driven mostly by the same known customers. That strengthens the reading that the 2026 issue is not lack of commercial interest, but conversion quality. If part of current backlog slips to the right, the company has less room than it had a year ago to offset that with a broader and more diversified opportunity pool.
The product-line split does not really diversify the picture either. About 58% of pipeline sits in satellite solutions and supporting infrastructure, 25% in satellite services, and 17% in intelligence as a service. So the future opportunity pool is also relatively tilted toward the company's more complex deal type, not only toward straightforward extensions of existing service contracts. That supports the long-term case, but it is a less reliable short-term safety net for any 2026 recognition miss.
Bottom Line
The right way to read ImageSat's 2026 backlog is neither "fully locked" nor "fragile." There is a real contractual base here: $160.8 million of backlog, $78.4 million mapped into 2026, and a meaningful service backbone. That is the part that gives the company a real floor for the coming year.
But above that floor sits a much more sensitive layer. Concentration in Customer A and Chile is severe, and the Chile side has already shown that signing the contract does not settle the timing question. Delays, a revised work plan, and one canceled scope have already happened. So 2026 is not a demand-proof year. It is a conversion-proof year. If the company shows that the first half absorbs the deferred milestones and that Chile execution stays on the updated track without another slip, the market's reading of backlog should improve materially. If not, the question will shift from how much backlog exists to how much of it is truly available for timely revenue recognition.
Disclosure: Deep TASE analyses are general informational, research, and commentary content only. They do not constitute investment advice, investment marketing, a recommendation, or an offer to buy, sell, or hold any security, and are not tailored to any reader's personal circumstances.
The author, site owner, or related parties may hold, buy, sell, or otherwise trade securities or financial instruments related to the companies discussed, before or after publication, without prior notice and without any obligation to update the analysis. Publication of an analysis should not be read as a statement that any position does or does not exist.
The analysis may contain errors, omissions, or information that changes after publication. Readers should review official filings and primary sources before making decisions.