Eastern Pacific Shipping Expands Orca AI Across More Vessels

Eastern Pacific Shipping is scaling Orca AI across more vessels after a pilot cut close encounters and widened passing distances.

Eastern Pacific Shipping (EPS) is expanding its deployment of Orca AI’s maritime operations platform after a five-vessel pilot recorded reductions in close encounters and increases in vessel passing distances.

The Singapore-based shipowner plans to install the technology on dozens of additional vessels following more than 130,000 nautical miles of operational testing. The move takes the project from a limited trial into broader fleet deployment, providing a larger test of how AI-assisted navigation tools perform across day-to-day shipping operations.

During the pilot, Orca AI reported a 48% reduction in close-encounter events per 1,000 nautical miles across participating vessels. Average minimum passing distance increased by 10%.

One vessel, the Pacific Sentinel, recorded an 84% reduction in open-water close encounters and a 50% increase in minimum passing distance during the six months following installation.

EPS Moves AI Navigation From Trial to Wider Deployment

The decision to expand the system is significant because maritime technology pilots do not always progress to fleet-wide use. Operators need to establish whether a new platform delivers practical benefits for bridge teams while also producing data that can support safety, training and operational decisions ashore.

EPS Technical Director Sachin Saharawat said the pilot was designed around supporting masters and bridge teams with additional situational awareness rather than replacing their professional judgment. EPS also used operational data from the trial to monitor changes in navigational risk over time.

Those findings have given the company sufficient confidence to broaden the deployment.

EPS manages a fleet of more than 350 vessels spanning containerships, vehicle carriers, dry bulk ships, gas carriers and tankers. The companies have not disclosed the exact number of ships included in the next phase, beyond stating that the technology will be rolled out across dozens of additional vessels.

The larger deployment should offer a more representative test than the original five-vessel pilot. Different ship types, routes, traffic conditions and crew practices can all influence navigational risk, making consistency across a larger fleet an important measure of the technology’s effectiveness.

Orca AI’s system combines onboard computer vision with shoreside fleet analytics. Its SeaPod hardware is installed on the bridge and is designed to detect and classify vessels and other objects in the surrounding environment.

That includes some targets that may not be broadcasting Automatic Identification System signals, such as fishing vessels or navigational objects. The additional visual data is intended to complement conventional bridge equipment and watchkeeping, particularly in congested waters or reduced visibility.

Fleet Data Could Become as Important as Bridge Alerts

The expansion also highlights a broader use case for AI in shipping: turning individual navigational events into fleet-level operational data.

Orca AI’s FleetView software allows shoreside teams to review events across multiple vessels and identify recurring patterns in navigational behaviour. For EPS, that information can be used to support crew training, performance reviews and broader safety analysis.

This type of fleet visibility could give operators a more structured way to assess navigational risk than relying solely on incident reports or individual voyage reviews. It may also help companies identify emerging trends before they develop into more serious safety issues.

Orca AI CEO and co-founder Yarden Gross has said the EPS deployment shows how AI-based situational awareness can support existing seamanship and decision-making both onboard and ashore.

The company says more than 1,600 vessels have been booked to use its platform, with customers including MSC, Seaspan, NYK and Maran Tankers.

However, the next stage of the EPS rollout will be the more meaningful test. Results from five vessels cannot automatically be applied across an entire fleet. Traffic density, voyage profiles, operating regions, crew behaviour and baseline safety performance can all affect close-encounter statistics.

If similar reductions can be maintained across a broader group of vessels, the deployment could strengthen the case for AI-based navigation systems as a routine fleet-management tool rather than an isolated technology experiment.

For shipowners, the value may ultimately extend beyond individual bridge alerts. The ability to track and compare navigational risk across vessels could become an increasingly important part of safety management as fleets generate more consistent operational data.