Idomoo: Is AI Already In ARR, Or Still Mostly In Backlog And Promises
By the end of 2025, AI is already inside Idomoo's ARR, but at $1.8 million it is still less than 9% of total ARR. This is no longer just backlog and promises, but it is not yet a broad enough revenue layer to prove that the new commercialization engine has fully settled.
What This Follow-Up Is Actually Testing
The main article already framed the central tension at Idomoo: the AI engine is moving faster than the balance sheet. This follow-up isolates one narrower question inside that broader thesis: how much of the AI story is already sitting inside recurring revenue, and how much still lives mainly in backlog, signed deals and management promises.
The short answer is that AI has already entered ARR. By year-end 2025, AI ARR reached $1.8 million, versus only $0.3 million a year earlier. That is real recurring base, not just demos and not just presentation material. The commercial signal set is broader too: around 25 Lucas AI deals were signed during 2025, the first AI customer doubled its ARR contribution to about $0.6 million, and the first AWS Marketplace deal carries about $1.17 million of ARR.
But this is still not maturity. AI ARR is less than 9% of total ARR of $20.3 million. The company also defines ARR in a way that includes signed customer orders and expansions that have not yet been activated, so the metric is much stronger than a vague sales promise, but it is still not the same thing as revenue already recognized. The right reading is that AI has crossed from promise into early commercialization, but it has not yet built a wide, stable and diversified enough revenue layer to fully reset the quality of the story.
What Is Already Inside ARR
The most important point is not only that AI ARR increased, but that the increase is now tied to tangible contracts. The company's first AI customer, an international crowdfunding platform, renewed and expanded its Lucas Enterprise AI Video usage in October 2025 for 12 months, with ARR of about $0.6 million versus about $0.27 million in the original contract. That is not a trial order. It is an expansion that doubles an existing recurring base.
The AWS thread also moved from strategy to actual numbers. The first AWS Marketplace deal, signed on December 31, 2025 with an existing US mortgage customer, generates about $1.17 million of ARR, roughly 30% above the earlier agreements with that same customer and the business it acquired. That is a meaningful commercial signal because it proves the channel exists, but it also shows the real state of play at this stage: the first proof point came from expanding an existing relationship, not from opening a broad new-customer engine.
Breadth of adoption is also starting to show. Around 25 Lucas AI deals were signed during 2025. That means the product is no longer sitting on the edge of the company. It has entered the selling layer. What is still missing is evidence of how many of those deals become meaningful recurring revenue within a reasonable time frame, and how many remain a useful add-on for existing accounts.
Why This Is Still Not "Clean" ARR
The core of the issue sits in the ARR definition itself. The company explains that ARR includes recurring monthly revenue in a representative month, plus expected revenue from signed orders and agreements with new customers that have not yet been activated, and plus signed expansions for existing customers that have not yet been activated. At the same time, it explicitly says ARR is not a revenue forecast, not deferred revenue and not a substitute for those metrics.
That is the critical nuance. At Idomoo, ARR is much stronger than a generic promise because it rests on signed orders and agreements. But it still sits one step before full revenue recognition. So even the $1.8 million of AI ARR clearly signals real commercialization, while not necessarily meaning that the full amount has already flowed through the income statement.
| Layer | What it actually means | What it should not be confused with |
|---|---|---|
| Total ARR of $20.3 million, including AI ARR of $1.8 million | Annualized recurring base, including signed orders and expansions that are not yet activated | Not recognized revenue, not deferred revenue, and not a substitute for either |
| Backlog of $15.7 million at the end of 2025 | Binding minimum commitments not yet recognized, with visibility mainly into 2026 | Not proof that AI already dominates backlog, because the company does not split AI backlog from the core backlog |
| Management expectation for about $6 million of AI bookings in 2026, versus about $2 million in 2025 | A sales target that points to future signing pace | Not ARR, not revenue, and not visibility that has already settled |
Backlog itself reinforces the cautious reading. At the end of 2025, backlog stood at $15.7 million, of which about $12.1 million is expected to be recognized during 2026. That is a good starting base for the coming year, but it is not labeled as AI, and the company also notes that a meaningful share of quarterly contracts is usually signed in the final 10 days of the quarter. The fact that backlog already stood at $13.8 million near the report date is a reminder of how quickly the picture can shift between signing, activation and recognition.
Commercialization Quality Still Rests On The Existing Base
The most encouraging part of commercialization quality is that AI is not being sold into a vacuum. It is being sold into an established and familiar customer base. Idomoo has 165 active customers, 63 of them with tenure of five years or more, versus 54 at the end of 2024. That is real ground for upsell. Geographic spread is also reasonably broad, with 102 customers in North America and 63 in Europe and Asia, and the company states that it has no dependency on a single customer.
But commercialization quality is still not truly broad. The top 10 customers already account for 46% of ARR, up from 41% a year earlier. At the same time, 39 ARR customers did not renew in 2025, the same count as in 2024, but the dollar value lost rose to $2.48 million from $1.84 million. In other words, the company returned to ARR growth, but it did so against a more expensive erosion of the existing base.
The sector mix makes the same point. In 2025, 56% of revenue came from financial services and another 11% from telecom, while the company serves 4 of the 5 largest US banks. That is an important commercial advantage, because those kinds of enterprise customers do not bring in AI solutions lightly. But it also means the new engine still rests mainly on deepening heavy enterprise accounts rather than on rapid horizontal spread.
That is also why the first AWS Marketplace deal matters, but is not enough on its own. It proves the channel works, yet for now it improved commercialization of an existing customer. For the market to start treating this as a genuine new growth engine, Idomoo will have to show that the model repeats across additional customers, without every AI gain arriving with even higher concentration.
AWS And Lucas: The Signal Is Positive, Repeatability Is The Test
The Lucas thread and the AWS thread meet exactly at the point that makes this follow-up interesting. Lucas gives the company a product layer that can enter more deals. AWS adds a channel, technical support and implementation leverage through Amazon Q and AWS Marketplace. Those two things together can shorten the path from an AI product to ARR.
But as of the end of 2025, this is still an early proof stage. The AWS agreement gives the company preferred-partner status, technical support and joint marketing efforts. Those conditions can materially improve commercialization quality. They do not create ARR by themselves. ARR appears only when a transaction is signed. So far, the first transaction through that channel is one deal, with one existing customer, signed on the last day of the year.
That is why the 2026 test will not be whether the company has an AI narrative. It already does. The test will be whether management's expectation for about $6 million of AI bookings in 2026, versus about $2 million in 2025, translates into three hard outcomes:
- a material increase in AI ARR beyond the current $1.8 million layer
- more than one partner-driven deal that actually sits in ARR, not only point expansions of existing customers
- a halt in the rise of dollar churn and customer concentration within the ARR base
Until that happens, the right sentence about Idomoo is not that AI still lives only in promises. It is that AI is already inside ARR, but still mainly as a new floor built on top of an older customer base, not yet as a broad and independent commercialization engine.
Conclusion
Bottom line: AI is already inside Idomoo's ARR, but for now it sits there more as a deepening of the existing customer base than as proof that a new commercialization engine has fully stabilized.
This is no longer just backlog and promises. $1.8 million of AI ARR, the expansion of the first Lucas customer, and the first AWS Marketplace deal are real commercial facts. But this is also not yet "high-quality" ARR in the broader sense. The metric includes signed expansions that are not yet activated, a meaningful part of the story still rests on large existing customers, and concentration among the top 10 customers has increased.
So the right reading is an in-between one: Idomoo has already shown that it has an AI product it can sell inside its customer base. It has not yet shown that this sale is broad enough, repeatable enough and diversified enough to argue that the AI layer now sits deep inside the core quality of revenue.
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.