A Tel Aviv-based startup says it has developed a way for autonomous underwater vessels to communicate over long distances without surfacing, addressing a longstanding challenge in naval and maritime operations. Skana Robotics said its approach uses AI-driven coordination rather than the large language models currently popular in the technology sector.
A communication challenge beneath the surface
Underwater autonomous vessels and robots are increasingly seen as useful tools for defence, surveillance, and infrastructure protection. However, communication has remained a limiting factor. Submersibles typically struggle to transmit information across large distances unless they surface, a step that can expose them to detection.
Skana Robotics said it has built a new capability into its fleet management software, SeaSphere, that allows groups of vessels to communicate with each other underwater while remaining submerged.
How SeaSphere coordinates vessels
According to Skana, the updated SeaSphere system allows vessels to share data and respond to signals received from other robots in the fleet. The software enables individual units to adjust their behaviour based on incoming information, while still operating toward a shared mission objective.
The company said this coordination allows fleets to adapt autonomously to changing conditions. Skana added that the same system could be applied to securing underwater infrastructure and maritime supply chains.
Idan Levy, co-founder and chief executive of Skana Robotics, told TechCrunch that communication remains one of the main obstacles in large-scale unmanned operations.
He said the company is focused on how hundreds of unmanned vessels can be deployed in a single operation, share data, and maintain communication both at the surface and underwater.
Older AI models for predictability
The research behind the new capability was led by Teddy Lazebnik, an AI scientist and professor at the University of Haifa. Lazebnik said the team deliberately avoided newer AI models and instead relied on older, more mathematically grounded algorithms.
He told TechCrunch that while newer AI systems can be more powerful, they are often less predictable. Lazebnik said the algorithms Skana chose offer greater explainability, predictability, and generality, even if they lack the performance gains or novelty associated with newer approaches.
Company background and next steps
Skana Robotics was founded in 2024 and emerged from stealth earlier this year. The company is currently targeting government agencies and commercial customers in Europe, where maritime security concerns have increased amid the war between Russia and Ukraine.
Levy said Skana is in discussions for a large government contract that it aims to secure by the end of the year. The company plans to release a commercial version of its software in 2026 and demonstrate its performance in operational environments.
Lazebnik said Skana wants to show that its software can manage complex manoeuvres at scale and handle real-world operations. He said the company plans to invite naval leaders in European countries to evaluate the system and assess its results directly.
Featured image credits: Wikimedia Commons
For more stories like it, click the +Follow button at the top of this page to follow us.