Analysis and buyer guides
25 articles
Three Quarters of Israeli Defence-AI Firms Have No Public Deployment Evidence
Of 50 Israeli defence-AI companies indexed, only 13 have a public source describing operational military use. Here is what the other 37 do and do not show.
Funding Stage Is Not Deployment Evidence
Israeli defence-tech raised over $1bn in 2025, but the index's best-funded entries sit at L2 and L0. Why funding stage tells a buyer nothing about fielding.
What "Battle-Tested" Means When Nobody Publishes the Test
"Battle-tested" is the most common phrase in Israeli defence marketing and the least verifiable. How the index grades a claim when the test is never published.
The Undisclosed-Customer Problem in Defence Contract Announcements
An undisclosed defence contract announcement is verifiable and nearly useless. How "a European NATO member" and "an APAC agency" shape 22 L2 grades in the index.
Three Categories Came Back Almost Bare
Three of twelve Israeli defence AI categories yielded one qualifying company each. Active protection, fire control and sustainment, and why startups are absent.
The English-Language Bias in Israeli Defence-Tech Research
All 50 index sources are English-language. Why that biases Israeli defence tech research downward, which grades it suppresses, and what would correct it.
Press Release or Reporting: Grading the Difference
Eight of 50 index grades rest on issuer-controlled releases and none of them is L3. How defence press release verification separates a wire item from reporting.
The 2023–2024 Cohort and Its Evidence Gap
Six Israeli defence tech startups founded in 2023 or 2024 appear in the index. None is graded L3. What that reveals about time to public deployment evidence.
Why the Israeli Primes Are Excluded From This Index
Elbit, Rafael and IAI reported roughly $22bn of 2025 revenue between them and are absent from this index. The inclusion rule, and the honest cost of applying it.
Why This Index Does Not Link Products to Named Strikes
The index records that a system was reported in use. It never links a product to a named strike. The rule, the reasons for it, and the case against it.
Counter-UAS Is the Most Crowded Category in Israeli Defence AI
Eleven of 50 index entries are counter-UAS, spread across every grade. Four of them sit at L0. The crowding, the arithmetic and what it means for buyers.
What Moves a Company From L0 to L3
Each grade in the index has one threshold and one document that crosses it. Here is exactly which published artefact moves a company from L0 to L1, L2 and L3.
What a Deployment Index Cannot Tell You
Six questions this index of 50 Israeli defence-AI companies cannot answer, stated plainly, with the specific entries that show where each limit bites.
How to Verify a Defence AI Vendor's Deployment Claims
A four-step method to verify defence vendor claims before you shortlist: parse the sentence, find the primary record, date it, and grade what actually remains.
Procurement Due Diligence for Autonomy Suppliers: A Checklist
Twenty-five checks across evidence, autonomy, export, data and sustainment: a defence procurement due diligence checklist for buying autonomy, not slideware.
Israeli Counter-Drone Companies With Publicly Announced Contracts
Six Israeli counter-drone contracts sit on the public record, from Robotican to Sentrycs. Each entry names the publisher and links to the source you can open.
The 13 Israeli Defence-AI Companies With Documented Operational Use
Thirteen Israeli military AI companies have a named public source reporting operational military use. Every entry lists the publisher and a link you can open.
Israeli Defence Export Controls: What a Foreign Buyer Should Check
An Israeli defence export licence is not one document but several, issued at different stages by two ministries. What a buyer should ask, and where to check.
Israeli Defence-Tech Landscape Maps Compared
Published Israeli defence tech landscape map projects count between 160 and 334 companies. What each measures, why the totals differ, and what all of them omit.
Detect, Track, Defeat: A Counter-UAS Vocabulary for Buyers
The counter-UAS detect, track, defeat sequence is doctrine, not marketing. What each stage means, who defines it, and which claims to test first.
Loitering Munitions and Attack Drones: Definitions That Matter
A loitering munition definition is not settled. Doctrine, arms control and humanitarian law each draw the line differently. Here is what each one says.
In the Loop, On the Loop, Out of the Loop
In the loop, on the loop, out of the loop: what human in the loop weapons policy says, who defines each term, and precisely where states disagree today.
Buying Military Simulation and Training AI
Military simulation procurement turns on three things: the LVC taxonomy, interoperability standards, and accreditation. Plus what the Israeli record shows.
TRL Is Not Deployment Evidence
A technology readiness level describes maturity in a test environment. Deployment evidence names who used the system. Why buyers should not confuse them.
Twenty Questions for a Defence AI Vendor's First Meeting
Twenty defence AI vendor questions for a first meeting, each with why it matters and what a weak answer sounds like. Built from public procurement guidance.
The full index
All 50 companies, graded by publicly documented deployment evidence.