Ondas Builds Regna AI on Palantir, With Economics Still Undisclosed
Ondas launched Regna AI as a wholly owned business using Palantir’s AIP for sovereign defense models. The technical plan is specific, but investment commitments and expected revenue were not disclosed.
Ondas has launched Regna AI, a new wholly owned business designed to let governments and companies train and deploy specialized artificial intelligence within infrastructure they control. Palantir will provide its Artificial Intelligence Platform, model-training infrastructure and forward-deployed engineering support.
The product addresses a concrete defense problem. Sensitive mission data may not be allowed to leave a government-controlled environment, while battlefield communications can be degraded or unavailable. Regna AI is intended to run customized models on local hardware across drones, stratospheric balloons, ground vehicles, surface vessels and command systems.
Ondas will be the first customer. The venture builds on the companies’ SkyWeaver C4I collaboration, which integrates surveillance from high-altitude platforms to the tactical edge. Beginning with intelligence, surveillance and reconnaissance, Regna plans to expand into security, critical infrastructure and industrial markets.
The sovereign model has a practical attraction. Governments can retain control of operational data and the models trained on it without spending the hundreds of millions or billions of dollars that Ondas says a frontier model can cost to build from scratch. Palantir supplies an existing software and deployment layer, while Ondas contributes autonomous systems and customer access.
The announcement does not yet establish a financial case. Ondas and Palantir did not disclose investment commitments, pricing, ownership changes beyond Ondas’s initial wholly owned position, or expected revenue. Ondas serving as the first customer demonstrates a use case, but it is not the same as an external commercial contract.
Execution will also be demanding. Models used for defense decisions need secure training data, auditable behavior, reliable operation at the edge and clear rules for human control. Deploying software across multiple hardware types creates integration and maintenance costs that a press release cannot quantify.
For Ondas, Regna AI could increase the software and recurring-service content attached to its autonomous platforms. For Palantir, it creates another route into sovereign and defense workloads. The value to shareholders depends on external customer wins, contract economics and the amount of capital Ondas must commit.
The launch is strategically credible because it extends an existing collaboration and names a first deployment. It remains financially preliminary because the terms that would convert the technical concept into revenue are missing.



