Defense technology group Ondas and data analytics firm Palantir Technologies are forming a new company, Regna AI, built to give governments and businesses a way to deploy artificial intelligence in sensitive operations without handing control of their data or models to outside providers. The venture addresses a problem that has grown more urgent as militaries and critical-infrastructure operators look to adopt AI: how to gain its advantages without exposing classified or operationally critical information to systems they do not fully control.
The appeal of commercial AI models is obvious, but so is the risk. Organizations operating under strict security requirements often cannot send their data to external cloud platforms, no matter how capable the underlying model. This is a concern that echoes broader debates in digital security circles, where even ordinary users weighing a fast vpn for private browsing must consider who controls the infrastructure carrying their traffic. Regna AI is designed around the same underlying principle, scaled to military and industrial stakes: customers should be able to run specialized models inside their own environments, including hardware deployed in the field, rather than depending on infrastructure managed elsewhere.
How the Venture Is Structured
Ondas will initially own Regna AI outright and serve as its first customer, while Palantir supplies the infrastructure needed to train and adapt models to specific operational needs. Palantir's software is already widely used by government and commercial clients to manage and analyze operational data, giving it an established role in environments where data sovereignty is non-negotiable. The two companies will work together on engineering tailored to how customers actually operate, rather than offering a generic model meant to serve every use case.
The company's initial focus is intelligence, surveillance and reconnaissance, with plans to expand into broader defense applications, security, critical infrastructure and industrial markets. A central use case involves autonomous systems operating where communications are unreliable or deliberately jammed. The models are meant to run on drones, stratospheric balloons, ground vehicles, surface vessels and command-and-control systems, helping operators assess evidence and weigh courses of action even when a link back to central servers cannot be guaranteed.
Human Oversight Remains the Default
Despite the autonomy built into these systems, the companies say any AI-generated recommendations for updating mission plans will remain subject to human review and authorization. That distinction matters: as militaries integrate AI into decision-making, the question of where human judgment sits in the loop has become one of the defining policy debates in defense technology, shaping both operational trust and the rules governments are drafting to oversee autonomous systems.
Built on Existing Infrastructure
Regna AI will draw on an existing collaboration between Ondas and Palantir involving SkyWeaver C4I, an open-architecture system built to integrate surveillance across platforms ranging from the stratosphere to the tactical edge. That groundwork gives the new venture an operational base for running models across varied hardware and degraded network conditions. Ondas, which employs roughly 1,000 people in Israel, has been expanding rapidly through acquisitions, having bought more than 15 defense technology businesses over the past two years, including three companies acquired last month for a combined $56 million, spanning drones, robotics, counter-drone systems and communications.
