Windward adds camera-based ship identification with Azimut.ai acquisition

Ships that stop transmitting are the target of a new camera-based sensing layer at Windward. The maritime AI company announced its acquisition of Azimut.ai on November 3, 2026, adding technology it says can identify vessels in real time at ports and maritime chokepoints, then connect those sightings with its existing intelligence.

Windward did not disclose the purchase price or explicitly confirm that the transaction had closed. It described the acquisition as its first step into ground- and ship-based sensors, extending a platform that already combines several sources of maritime data.

Connecting a sighting to a ship’s history

The company’s existing inputs include vessel-transmission data, satellite imagery and radio-frequency signals. Windward points to gaps in that picture: not every vessel transmits, and signals can be switched off, jammed or affected by external conditions. Weather can also obscure imagery. Cameras are intended to add another way to establish a vessel’s presence within their coverage.

Windward’s plan includes placing cameras at key thoroughfares and maritime chokepoints. Its stated aim goes beyond detecting a ship: visual identification would be combined with records of ownership, sanctions, cargo, behavior and risk already held in the platform. A single sighting would become an assessment that government or commercial customers could use, rather than a camera observation without that context.

Acquiring Azimut is our first step in building a proprietary visual intelligence layer that gives us full visibility at sea.

Ami Daniel, Windward CEO and co-founder

What Azimut’s cameras can contribute

Azimut’s product documentation describes software that works with existing cameras without requiring new hardware, according to Unite.AI. It can combine multiple cameras and sites into one maritime-awareness map. The software detects and classifies vessels in real time, while a recognition function uses visual features to identify ships and connect them with vessel details.

The documented functions also include calculating position, course and speed from video alone. The system learns behavior specific to a site over time, automates camera-scanning patterns and issues alerts for unusual activity. Those functions add movement and local behavioral context to the vessel identification Windward intends to combine with its other datasets.

Unite.AI also cited Azimut’s results from an Ashdod Port pilot: the company reported detecting and documenting 464 targets, with detection ranges of up to 7,000 meters using standard day cameras. These are company-reported results from that pilot, not a measurement of the combined Windward platform.

Vessel analytics, without the raw footage

Windward says customers receive no raw images or video footage. It describes the system as designed for privacy and security, with a remit to analyze vessel activity rather than identify or verify individuals. The customer-facing output is therefore vessel intelligence, even though cameras supply the underlying observations.

That output fits Windward’s existing processing structure. Multiple datasets feed a shared operational picture; its AI then applies behavioral pattern recognition, vessel verification and anomaly detection. The acquisition adds a visual input to that chain rather than replacing the other sources. Customer-specific capabilities are designed around operational needs, which Windward describes as the partnership layer of its platform.

Building the sensing layer also involves finding places to observe maritime traffic. Alongside the acquisition announcement, Daniel invited maritime stakeholders and others to become site partners. The invitation accompanies Windward’s plan to place cameras at key locations and bring a proprietary source of visual evidence into its intelligence platform.