What stands out to me isn't the laser tech, it's the reframe. EON isn't selling faster bandwidth, it's selling a failure mode nobody can correlate with the seabed. That's the actual moat, not speed, not cost.
One system going down shouldn't take everything else with it. Actually just wrote about this on my recent article, The AI Stack That Automates My Life
Latency — MEO bent-pipe is 2–2.8× slower than fiber, not faster. NY–London: 60–86 ms vs 30.6 ms. The vacuum speed advantage only accrues to a LEO mesh, which is a different architecture and one SpaceX already runs.
Redundancy — A cable carries 16–24 independent fiber pairs with mesh restoration. One satellite serves one route; its failure takes the route down. And the optical ground stations still need terrestrial fiber backhaul — the land-side dependency doesn't go away.
Cost — $25M per delivered Tbps optimistic, $228M realistic, vs ~$0.65M for a modern cable. Structural, not first-generation: fiber adds a line card, space adds a satellite. Cheaper launch doesn't close it.
Outages — Cloud blocks the link entirely; single-site availability is 50–80% against a 99.99% carrier standard. Add handover gaps, sun-in-view, radiation, and fixed, publishable ground sites that are softer targets than the cables this is meant to replace.
Most of the a16z news is compelling. This is an interesting concept. But I couldn't get past the annoying AI prose to read all of it: One prompt. Two buzzwords. Nothing original on the page.
If you don't have the hour to hand-craft this kind of message, then farm it out to an intern and buy them lunch. They would've done a better job conveying your message.
What stands out to me isn't the laser tech, it's the reframe. EON isn't selling faster bandwidth, it's selling a failure mode nobody can correlate with the seabed. That's the actual moat, not speed, not cost.
One system going down shouldn't take everything else with it. Actually just wrote about this on my recent article, The AI Stack That Automates My Life
Latency — MEO bent-pipe is 2–2.8× slower than fiber, not faster. NY–London: 60–86 ms vs 30.6 ms. The vacuum speed advantage only accrues to a LEO mesh, which is a different architecture and one SpaceX already runs.
Redundancy — A cable carries 16–24 independent fiber pairs with mesh restoration. One satellite serves one route; its failure takes the route down. And the optical ground stations still need terrestrial fiber backhaul — the land-side dependency doesn't go away.
Cost — $25M per delivered Tbps optimistic, $228M realistic, vs ~$0.65M for a modern cable. Structural, not first-generation: fiber adds a line card, space adds a satellite. Cheaper launch doesn't close it.
Outages — Cloud blocks the link entirely; single-site availability is 50–80% against a 99.99% carrier standard. Add handover gaps, sun-in-view, radiation, and fixed, publishable ground sites that are softer targets than the cables this is meant to replace.
Most of the a16z news is compelling. This is an interesting concept. But I couldn't get past the annoying AI prose to read all of it: One prompt. Two buzzwords. Nothing original on the page.
If you don't have the hour to hand-craft this kind of message, then farm it out to an intern and buy them lunch. They would've done a better job conveying your message.
Very intriguing tech. Best of luck and I'll keep watching for more.
EON sounds incredible. There is a dynamic ambient superconductor spec on my stack that is TRL 3/4. Maybe that could improve things somehow?