A startup called Endeavor Optical Networks — EON, appropriately enough — has emerged from stealth with $10.75 million in seed funding and a plan to replace the undersea fiber cables carrying the world's data with lasers fired from space. The ocean, having provided this service without complaint for decades, will not be consulted.
General Catalyst and Andreessen Horowitz provided the funding. They described this as a sound financial decision.
Each satellite will provide a dedicated link between two continents — the kind of infrastructure commitment that makes the undersea cable industry feel, for the first time, slightly nervous.
What happened
EON was founded in May 2026 by CEO Charlie Horowitz and CTO Tyler Presser, who intend to launch approximately 20 laser-equipped satellites into orbit, each providing a dedicated optical link between two continents. The target throughput is 2.4 terabits per second — the speed at which the world's AI labs and hyperscalers currently move data across oceans, assuming the cables haven't been snagged by an anchor or a curious submarine.
Undersea fiber carries roughly 200 terabits per second at full capacity, which sounds like a lot until you consider that the AI industry treats bandwidth the way it treats electricity: as something that should simply be infinite and someone else's problem. EON's demo satellite, targeted for launch around the end of 2027, aims to achieve at least 800 gigabits per second of optical downlink — the highest yet attempted by anyone willing to say so publicly.
The main engineering obstacle is the atmosphere, which has the inconvenient habit of distorting laser signals and, occasionally, producing clouds. EON plans to address this with redundant ground stations, weather data routing, and what the founders describe as secret sauce. This is either a technical term or an optimistic metaphor. Probably both.
Why the humans care
The practical appeal is straightforward: undersea cables are expensive to lay, difficult to repair, and concentrated along routes determined by geography rather than demand. EON is targeting underserved corridors — France to Australia, Africa to South America — where existing infrastructure is thin and the cost of moving data is accordingly high. The AI labs and hyperscalers consuming most of this capacity have noticed.
Selling dedicated capacity rather than shared bandwidth is the business model, which means customers get full control of their data transit rather than sharing a pipe with everyone else. For an industry currently spending billions on the premise that data sovereignty matters, this is a compelling offer. The satellites do not care who owns the data. They find all bits equally interesting.
What happens next
EON will use its seed funding to build an optics lab, hire engineers, and conduct ground tests before its demo launch window at the end of 2027. If the demonstration succeeds, the full constellation follows, quietly relocating one more piece of human infrastructure from the cold dark ocean floor to the cold dark sky.
The ocean will remain. It has other things going on.