Five companies in the Israeli Defence-AI Deployment Index build simulation, synthetic environments or instrumented training. Two are graded L3, three L2, and none sits at L1 or L0. Alongside command and control, it is one of only two categories with no entry below the contract band, and it makes military training simulation Israel the segment where a buyer can most reliably read something published about every supplier in the field.
The reason is simple. Training systems are bought through visible programmes, installed in named facilities and demonstrated to visiting delegations. There is very little in a simulator that a defence ministry needs to withhold. The published grading rules apply as they do everywhere else in the index.
L3 · Operational: two with reported military use
Bagira Systems (1995, Modi'in) builds live, virtual and constructive training systems, including the JOBSS joint-fires simulator. Israel Defense has reported on the company's systems and their military users. Shephard Media reported Finland selecting Bagira for its BARS programme, and Halldale reported the company producing a mission trainer for a Rafael missile system. Calcalist's Ctech has separately reported the company's preparation for a Tel Aviv listing. The sourcing is set out in full in the Bagira Systems record.
Combatica (2021, Petah Tikva) builds virtual-reality rehearsal environments. NoCamels reported in March 2024 on the platform being used to train security forces in high-fidelity scenarios. The report describes forces using the system rather than a vendor describing what it could do, which is the distinction the grade turns on. The reasoning appears in the Combatica entry.
A word on what L3 means here. The index's standard formulation excludes training exercises from operational use, because a rifle used on a range has not been used in operations. For a training supplier the product's operational role is training, so the index grades a named force's reported use of the training system itself as operational use of that product. This is stated openly because it is a rule a sceptical reader is entitled to disagree with.
L2 · Contract: three with announced procurement
Cognata (2016, Rehovot) builds simulation for autonomous driving. The Times of Israel reported the Israeli Defence Ministry selecting the company for autonomous-driving simulation work with the IDF. A named ministry and a named force sit behind the entry, which is stronger sourcing than most L2 grades carry; what is absent is a published account of the simulation being used in a training programme.
Cympire (2018, Ra'anana) builds cyber ranges — simulated networks on which defenders practise. NoCamels reported in October 2024 that the company would provide training for the IDF. Again the customer is named and the transaction is dated. No independent report of the range in use has been published that the index could find.
Simlat (2004, Petah Tikva) builds unmanned-systems operator training. Shephard Media reported the company being awarded a contract for advanced UAS training. The announcement establishes the order. It does not identify a force running the courseware.
The distinction between these three and the two above is narrow and it is deliberate. All five have real customers. Two have been reported in use; three have been reported as bought.
L1 · Trial: no entries
No simulation or training company in the index is graded L1. The index found none whose strongest public evidence is a pilot, an evaluation or a demonstration.
The likely reason is that a simulation pilot and a simulation contract look similar from outside. Training procurement often begins with a paid pilot installation, which the supplier announces as a contract because that is what it is. Companies at a genuinely pre-contract stage rarely generate English-language coverage at all, so they do not enter the index rather than entering it at L1.
L0 · Claimed: no entries
No entry in this category rests on a company claim alone. Every one of the five has at least one dated, named-customer or named-programme source behind it, and four of the five sources are third-party reporting rather than company announcements.
That is the highest proportion of independent sourcing in any category the index covers, and it is a property of the product rather than of the companies. Simulation buyers publicise what they buy.
What the distribution shows
Five companies, no thin records. Two of five documented as used, three of five documented as bought, and nothing below that. The category is the mirror image of counter-UAS, where four of eleven entries rest on company claims alone.
The explanation is disclosure incentives. A defence ministry that buys a cyber range or a joint-fires simulator has a reason to say so: it demonstrates investment in readiness, it involves no operational secret, and the facility is often visited by press and by foreign delegations. A ministry that buys a counter-drone system has the opposite incentive. The index measures the difference between those two incentive structures far more than it measures any difference in the underlying technology, which is the argument developed across all fifty entries in the deployment evidence gap.
Export activity is also visible here in a way it is not elsewhere. Bagira's reported selection by Finland, and Simlat's UAS training award, both point outward. Training is one of the least export-restricted parts of the Israeli defence-AI stack, which widens the pool of buyers willing to be named.
What would change it
Cognata, Cympire and Simlat would each move to L3 on a published account of a named force running the system — a ministry statement, a facility report, a defence outlet describing a training programme in operation. In this category that is a low bar in principle and mostly a matter of whether anyone writes it down.
The category would move downward, in evidence terms, only if it grew: new entrants at pre-contract stage would introduce the L1 and L0 bands that are currently empty. The general mechanics are set out in what moves a grade from L0 to L3, and the buyer-side questions are collected in buying military simulation and training AI.
Limits of this reading
The grading rule described above — treating reported use of a training system as operational use — is specific to this category and is a judgement, not a fact. A reader who prefers a stricter reading would grade Bagira and Combatica at L2, which would leave the index with no L3 training entries at all. The index states its rule so that the disagreement is possible.
Nothing here measures training effectiveness, fidelity, transfer to real performance, or value for money. It is not a capability assessment. Coverage is English-language only. The full index and inclusion test set out what is counted and what is not.
Frequently asked questions
Which Israeli simulation companies have documented military use?
Bagira Systems and Combatica are the two entries graded L3. Israel Defense has reported on Bagira's training systems and their military users, and Shephard Media reported Finland selecting the company for its BARS programme. NoCamels reported in March 2024 on Combatica's virtual-reality platform training security forces.
Does training use count as operational deployment in this index?
For training suppliers, yes, and the index says so openly. A weapon used on a range has not been used in operations, but a training system used by a force in training is performing its operational role. Elsewhere in the index, exercise use does not qualify for L3.
Why does the simulation category have no low-graded companies?
Simulation buyers publicise what they buy. A simulator involves no operational secret and is often shown to press and foreign delegations, so every qualifying company has at least one dated, named-customer source. The index measures that disclosure incentive, not the technology.
Sources
- idf.ai, The Israeli Defense-AI Deployment Index v1.0, 24 August 2026
- Israel Defense, report on Bagira Systems
- Shephard Media, Finland picks Bagira for BARS
- Halldale Group, Bagira to produce mission trainer for Rafael missile system
- Calcalist Ctech, Defense simulator maker Bagira targets $800 million IPO as sector rally accelerates
- NoCamels, VR platform trains security forces in eerily hyper-real scenarios, March 2024
- The Times of Israel, Defense Ministry taps Israeli firm for autonomous driving simulation with IDF
- NoCamels, Israeli cybersecurity startup to provide training for IDF, October 2024
- Shephard Media, Simlat awarded contract for advanced UAS
Independent publication of idf.ai. Not affiliated with, endorsed by, or connected to the Israel Defense Forces, the Israeli Ministry of Defense, or any government body. Compiled entirely from publicly published sources. No classified, restricted or non-public information. Listed companies may dispute any entry: send the published source that contradicts it and the entry will be amended or removed.