A UK-backed robotics start-up that promises to replace aging offshore vessels and crews with always-on underwater machines has emerged from obscurity with $5m (£3.95m) in pre-seed funding, signaling new investor interest in so-called “physical AI” plays aimed at the world’s most stubborn analogue industries.
Founded in 2025 by former NASA and ETH Zurich engineers, Bubble Robotics has secured rounds from Episode 1 Ventures, Asterion Ventures and Norrsken Evolve after being founded by London-based talent investor Entrepreneur First. The company already has more than $4 million worth of signed letters of intent in offshore wind, subsea infrastructure and maritime security, suggesting commercial traction is well beyond the typical pre-seed playbook.
The pitch is straightforward, if ambitious. Today, inspecting an offshore wind turbine, a buried data cable or a section of piping on the seabed typically requires a chartered vessel, a specialist crew and a daily bill that can run up to $100,000. According to Bubble’s founders, between 80 and 90 percent of these costs are related to the boat and the people on it, rather than the inspection itself.
“By eliminating this dependency, we enable a significant transformation in cost, security and operational frequency,” said Jean Crosetti, CEO and co-founder. “What was once episodic becomes continuous.”
The plan is to forego ship-based missions altogether and instead deploy fleets of resident autonomous robots that live at sea for months at a time, continually inspecting, monitoring and collecting data without human intervention. Crosetti compares the model to the satellite constellations that have transformed Earth observation over the last decade, pointing only down into the water column rather than up into the atmosphere.
The timing reflects a broader turning point. Cheaper edge computing, more powerful on-device AI, and the rapid expansion of low-Earth satellite connectivity have made sustained unmanned operations technically possible in a way that was not even three years ago. The macroeconomic appeal is equally significant: the offshore energy sector alone is expected to require an additional 600,000 workers by 2030, a deficit that no graduate program can meet in a timely manner.
Bubble sells its capabilities on a robotics-as-a-service basis, sparing its customers the upfront capital expenditures and offshore mobilization costs that have traditionally excluded smaller operators from high-frequency inspection systems. Target use cases include inspection of wind turbine foundations, cables, pipes and subsea structures; Benthos mapping, photogrammetry and biofouling monitoring for climate and biodiversity clients; and mine countermeasures, unexploded ordnance detection and continuous surveillance for defense and maritime security buyers.
The last category is becoming increasingly important. Recent incidents involving undersea data cables in the Baltic and North Seas have raised concerns about the security of underwater infrastructure for European governments and NATO and highlighted how little of it is still monitored. Persistent autonomous systems provide a way to maintain a continuous presence around sensitive assets without tying up scarce naval resources.
Alice Bentinck, co-founder of Entrepreneur First, said the founders stood out from the first moment they met at one of the company’s kick-off weekends. “Patricia and Jean formed a team based on a shared belief and complementary skills: Patricia has world-class technical credibility in robotics, Jean has unusual commercial instincts and intensity. Their pace of iteration throughout the program and strong customer obsession make Bubble Robotics a company to watch closely.”
For the broader SME ecosystem, the emergence of Bubble is a useful data point. This suggests that capital is still flowing into deep tech startups with credible commercial traction, even if more speculative AI is cool, and that the long-promised convergence of robotics, AI and connectivity is finally producing companies with connected revenue lines, not just demos.
