Anzu Robotics

Anzu Robotics

Anzu Robotics is an Austin, Texas company founded in 2023 by Randall Warnas, who previously helped launch DJI Enterprise, worked on thermal imaging partnerships at FLIR, and briefly served as chief executive of Autel Robotics. It began operations in 2024 with the Raptor series, an enterprise drone line positioned as an alternative to buying directly from DJI.

The Licensing Model

The Raptor was built under a technology licensing agreement with DJI, using the core design and internal chipset of the Mavic 3 platform, with manufacturing in Malaysia. Anzu’s stated terms were that no royalties flow back to DJI, there is no shared ownership of Anzu, and no customer data is reported to DJI. Software was developed independently.

That arrangement placed the Raptor in an unusual regulatory position. Because the aircraft is not manufactured in a covered foreign country and its parts largely come from Malaysia, Anzu argued it met NDAA supply chain standards. It is not, however, NDAA-compliant in the sense that federal buyers use the term, and Anzu has acknowledged that components such as the Raptor-T thermal sensor are sourced from China. Buyers subject to federal procurement rules should verify the position rather than assume it.

Corporate Status

XTI Aerospace acquired Anzu Robotics together with Drone Nerds on 10 November 2025 for $40 million, making Anzu the manufacturing arm of XTI’s drone business.

In February 2026 Anzu announced that the Raptor series was no longer available for purchase, citing persistent component shortages it had been unable to resolve. Demand late in 2025, driven partly by buyers reacting to NDAA 2025, depleted stock earlier than expected and disrupted manufacturing timing. The company said it would shift focus to a next-generation product built for current commercial and regulatory requirements.

Why It Matters

Anzu was the clearest test of a specific idea: that you could licence Chinese drone technology, manufacture it outside China, and sell it to buyers who cannot or will not buy DJI. The model worked commercially — demand outstripped supply — and then broke on the same supply chain the strategy was meant to route around.

That is the more useful lesson than the licensing structure itself. Relocating final assembly does not relocate component dependency, and component dependency is what ended the Raptor. Whether the next-generation product solves that, or simply moves the same problem, is the thing to watch.