Sub-Wavelength Precision at Deployment Scale

Reconfigurable intelligent surfaces promise to reshape the 6G radio environment — but only if the network can steer thousands of elements without drowning the air interface in pilot overhead. ZeroPilot is building a control-plane architecture engineered for carrier-grade beam precision at scale, designed to drive in-band pilot dependence toward near-zero rather than letting it grow with element count. We treat the surface's hardware behavior as a structured, physically-grounded prior — the foundation on which precise, predictable beam control can be built.

Predictive Control — Not Predictive AI

ZeroPilot is built on model-based predictive control over a compact, physically-grounded state — not machine learning. By declaring how a surface actually behaves, our control plane anticipates geometry changes and is architected so multi-surface handovers complete before the user notices, rather than reacting after performance degrades. Critically, this proprietary control plane operates above an open, non-normative hardware-behavior interface — the Phase API — which we are contributing to the RIS community. 

The open interface declares the hardware truth; ZeroPilot exploits it.

Orthogonal to SotA — By Design

Most RIS research races to optimize channel estimation and beamforming algorithms. ZeroPilot takes an orthogonal path: instead of competing with those methods, we give them a better starting point — a declared hardware prior that reduces what they must rediscover at runtime. The result is a control-plane architecture in active development, designed to complement rather than replace state-of-the-art signal processing, and grounded in an open interface specification any vendor, lab, or operator can adopt royalty-free.

Four grand challenges. One architectural insight.  A control-plane architecture for 6G RIS, in active development. Multiple patents pending