The Pulse
Ondas and Palantir Form Regna AI for Sovereign Models
Ondas and Palantir announced Regna AI on October 12, 2026, to develop fine-tuned AI models for governments and enterprises. Ondas says it will be the new company’s first customer, starting with intelligence, surveillance and reconnaissance
AI.info Team · Written with AI ·
A new company built around mission-specific models
Ondas and Palantir announced Regna AI on October 12, 2026, describing it as a standalone private company that will develop and deploy fine-tuned AI models inside customers’ own environments. Ondas will initially own the company and serve as its first customer; Palantir will provide model-training infrastructure and support delivery. The announcement did not disclose financial terms, a customer list beyond Ondas, or a timetable for a first external deployment.
Regna AI’s initial focus is intelligence, surveillance and reconnaissance, or ISR. Ondas says the models are intended to run on edge hardware, including drones, stratospheric balloons, ground vehicles and surface vessels, as well as command-and-control systems. The target environments include conditions with denied, degraded, intermittent or limited connectivity, where systems may not be able to rely on a continuous link to a central network.
Human review remains part of the mission workflow
The company says its models will help operators assess evidence, compare options and propose updates to mission plans. Those proposals, according to the announcement, are intended for human review and authorization rather than automatic execution. The release does not provide performance results, describe specific model architectures, or identify which open-weight models Regna AI plans to adapt.
Ondas Chairman and Chief Executive Eric Brock framed the venture around customer control of models and data. “No government should have to build its AI strategy on models or data infrastructure it does not control,” Brock said. He described Regna AI as combining open-weight models, customer-controlled mission data and Palantir’s operational platform.
Palantir supplies the training platform
Regna AI will be built on Palantir’s Artificial Intelligence Platform, or AIP. The companies also plan to use forward-deployed engineering to adapt the software to customer workflows and requirements. That approach puts Palantir in the infrastructure and implementation role, while Regna AI is presented as the company working with government, business and institutional customers.
Palantir Chief Executive Alex Karp argued that models tuned on operational knowledge can reduce reliance on general-purpose frontier systems. “Sovereign AI, trained on mission knowledge and tradecraft, outperforms frontier models while removing dependence on them,” Karp said. That is Palantir’s claim; the announcement includes no benchmark data comparing Regna AI models with frontier systems.
SkyWeaver is the starting point, not a deployment report
The new company builds on Ondas and Palantir’s existing work on SkyWeaver C4I, which the announcement describes as an open-architecture capability for integrating surveillance from the stratosphere to the tactical edge. Ondas says embedded agentic AI and edge processing in that program provide an operational base for Regna AI’s models in multi-domain settings. The release does not say that Regna AI has already completed a deployment or report measured results from operational use.
Ondas says the partners intend to support initial customer deployments over the coming months, beginning with ISR and later considering security, critical infrastructure and industrial markets. Its announcement argues that tailored models could avoid the expense of building a frontier model from scratch, but gives no product pricing or cost comparison for Regna AI. For now, the concrete commitment is narrower: Ondas is the first customer, Palantir supplies training infrastructure and support, and the venture says its proposed mission-plan changes will remain subject to human authorization.