Inside FinRAG
The fastest route from New York to London may be 550 km up
Light is about a third slower in glass than in a vacuum. That simple fact means satellites linked by lasers could, in theory, beat undersea fiber between New York and London by around 12 milliseconds. Here is the physics, the evidence and the catch.
In this article
The shortest path between two cities is a straight line under the ocean. The fastest path might not be.
In theory, a signal sent up 550 km to a satellite, passed across the Atlantic by laser and beamed down in London could beat the best undersea cable by about 12 milliseconds per round trip. That is the conclusion of a 2020 analysis by NSR, now part of Analysys Mason.
Twelve milliseconds sounds small. For the fastest trading firms, it is an eternity.
The race today: 59 ms
The benchmark is fiber. According to NSR, the fastest commercial route between New York's NY4 and London's LD4 data centers runs just under 59 ms round trip. Typical commercial undersea routes measure closer to 70 ms.
That best route already hugs the great circle, the shortest path over the Earth's surface, about 5,570 km. Cable can barely get straighter. So the limit is no longer the route. It is the glass.
The physics: light is about a third slower in glass
Light travels about 300,000 km per second in a vacuum. Inside optical fiber it slows to roughly 200,000 km per second, because glass has a refractive index of around 1.5.
Run the numbers on the New York to London great circle:
| Medium | Speed | Round trip (5,570 km each way) |
|---|---|---|
| Vacuum (or near-vacuum in orbit) | ~300,000 km/s | ~37 ms |
| Glass fiber | ~200,000 km/s | ~55 ms |
Even a perfect, perfectly straight fiber cannot beat about 55 ms. NSR puts the theoretical fiber floor at the same figure.
A satellite path is longer: up 550 km, across, and down again. But almost all of it runs through near-vacuum. The extra distance costs less than the slowdown in glass. In NSR's model, fiber's propagation delay is about 28% higher than a laser-linked satellite route at 550 km.
Why the lasers matter
The trick only works if satellites pass the signal to each other in space. NSR's model assumes inter-satellite links. Without them, the signal would have to bounce up and down between ground stations, and the advantage shrinks or vanishes. There is nowhere in the mid-Atlantic to put those ground stations anyway.
That hardware now exists at scale. In a 2024 talk reported by Advanced Television, a SpaceX engineer said Starlink runs more than 9,000 space lasers that move over 42 petabytes of data a day, with link uptime above 99% and some links spanning 5,400 km.
The space network is real. Whether anyone trades over it is another question. We found no public confirmation that any trading firm routes orders over satellite laser links.
The precedent: traders already go through the air
This is not as far-fetched as it sounds. Trading firms already beat fiber across oceans, using radio.
Shortwave signals bounce off the ionosphere and cross the Atlantic without any cable. Independent tester STAC measured one Chicago to London link by Raft Technologies at a mean of 24.8 ms one way. By our estimate, that is well under what any fiber could physically achieve on that route.
In 2018, Bloomberg traced shortwave antenna sites near Chicago, one linked in county records to a joint venture of Jump Trading and Virtu Financial.
The catch with shortwave is capacity: kilobits per second, enough for a few critical signals, but not full market data. Satellites could, in principle, offer far more bandwidth and the same advantage of not traveling through glass.
The catch
NSR is clear that its 12 ms figure is a theoretical benchmark. Several things stand between theory and a live trading link:
- Jitter. Satellites move constantly, so the path length changes. Trading firms care about consistent latency, not just average latency.
- Processing delay. Every receive and retransmit adds time unless equipment runs in cut-through mode.
- Capacity and cost. Low-orbit links offer far less bandwidth than undersea cable, and dedicated capacity would not be cheap.
- Reliability. Weather affects the ground links, and a trading route needs to stay up when it matters most.
What this means for trading infrastructure
Our reading: the latency race has run into physics, and physics points up.
- Routes are close to optimal. The best cables already follow the great circle, so remaining gains come from the medium, not the path.
- Air and vacuum beat glass. Microwave towers proved it on land, and shortwave proved it over oceans.
- Orbit is the next medium. The laser network exists. The open questions are stability, capacity and price.
- Watch for the first confirmed user. When a trading firm or network provider confirms a satellite route, the transatlantic race will have moved off the seabed.
Would you route orders through space to save 12 milliseconds?
References
- SatNews: LEO Speed, When Milliseconds are Worth Millions (NSR insight), October 14, 2020
- STAC Reports: Latency and throughput of Raft Chicago to London links
- Data Center Knowledge (Bloomberg): HFT traders dust off 19th century tool, June 2018
- Advanced Television: SpaceX reveals Starlink laser capacity, February 2, 2024
Speeds and round-trip times in the table are our own rounded calculations. NSR's satellite figures are theoretical and propagation-only. FinRAG.tech has no connection to the firms mentioned. This article is commentary, not investment advice.
Worth sharing?
Pass this article along to someone who would find it useful.
The work continues in the studio.
More build notes, walkthroughs, and perspectives from the FinRAG team.
Explore all stories
