Picture two satellites a few thousand kilometers apart, both moving at roughly seven kilometers per second along slightly different tracks. Now imagine pointing a pencil-thin laser beam from one to the other, holding it steady on a target the apparent size of a basketball, and pushing tens of gigabits per second through it. That is an optical intersatellite link in operation, and a growing fraction of the data flowing through space is already moving this way.

An optical intersatellite link, or OISL, is a free-space laser communication channel between two spacecraft. Instead of using radio waves to talk to a ground station and bouncing data back up to another satellite, two satellites talk directly to each other with modulated light. The result is a space-based network: traffic can hop from satellite to satellite without ever touching the ground until it reaches a destination customer.

The Technical Bits

The carrier is a laser, almost always operating in the near-infrared. The most common wavelength is 1550 nanometers - the same band that fiber-optic networks on Earth use, which means terminals can borrow a deep ecosystem of telecom-grade modems, amplifiers, and components. Some systems run at 1064 nm, which has historical roots in older defense and science programs.

Demonstrated data rates per terminal range from around 1.8 Gbps on early operational systems to more than 100 Gbps on advanced research payloads, with link distances anywhere from a few hundred kilometers to several thousand. There is no atmosphere between the two satellites, so the beam does not scintillate, scatter, or get absorbed. In a vacuum, light is a remarkably well-behaved messenger.

The hard part is pointing. To close a link, the transmitting terminal has to aim its beam at the receiver to within a fraction of a milliarcsecond and hold it there while both spacecraft move, vibrate, flex thermally, and slowly drift in their orbits. A typical terminal stack includes a coarse pointing assembly (a gimbal that swings the telescope toward the rough direction of the peer), a fine steering mirror inside the optical path that nudges the beam in microradian steps, a beacon and tracking sensor that detects the partner’s signal, an 8 to 15 centimeter telescope, the laser source itself, and the modem that turns bits into modulated light.

Acquisition is a choreography. Coarse pointing slews to the predicted location, the beacon laser sweeps a small area until the partner sees it, fine pointing locks on, and the high-rate data link comes up. Start to finish, this typically takes seconds to tens of seconds. If the link drops - a vibration spike, a thermal transient, a missed pointing update - the whole sequence repeats.

Why It Matters

The radio spectrum is a finite, regulated, increasingly crowded resource. Every gigahertz of bandwidth has to be coordinated through the ITU and national regulators, and crosslinks between satellites compete with downlinks, uplinks, and every other RF user on Earth. Traditional RF intersatellite links typically run somewhere between 50 and 200 Mbps, and going faster means either bigger antennas or rarer spectrum allocations. Optical links sidestep almost all of that. Photons need no spectrum license, and a single terminal can carry an order of magnitude more data than its RF equivalent.

The narrow beam also has security properties that matter to defense planners. An optical signal spreads only a few meters across by the time it reaches a peer satellite, which means an adversary trying to intercept the link would need to physically place a receiver inside that beam - extremely difficult in deep space. The same narrowness makes the link hard to detect (low probability of detection) and hard to jam (low probability of intercept). RF crosslinks, by contrast, scatter signals broadly enough that nearby satellites and ground sensors can passively listen.

The biggest architectural shift, though, is that OISLs eliminate the ground-station bottleneck. Without crosslinks, every satellite needs line-of-sight to a teleport to move data. With crosslinks, traffic can route through the constellation itself, crossing oceans and continents at the speed of light without ever descending into the atmosphere. This is the feature that makes proliferated low-Earth-orbit architectures actually work as networks rather than as collections of independent satellites. SpaceX’s Starlink runs on it. The Space Development Agency’s Transport Layer is built around it. OneWeb, Amazon’s Project Kuiper, and the planned Telesat Lightspeed all rely on it.

This connects directly to the medium-Earth-orbit story we explored earlier this week. MEO constellations have always had advantages over LEO for global coverage from fewer satellites, but moving large amounts of data between MEO assets has historically been hard. Optical crosslinks change that math, which is part of why the GPS III SV10 demonstration we covered yesterday matters so much: it is the first time a GPS-class navigation satellite has carried an OISL payload, and a successful demo opens the door to a future where the entire navigation constellation shares data directly without going through the ground network.

The Tradeoffs

Pointing tolerances at the milliarcsecond level are unforgiving. Thermal flexure of the spacecraft bus, mechanical vibrations from reaction wheels, even the slight torque of opening a solar array can knock a link off target. Designers spend enormous effort on isolation, calibration, and active control just to keep the beam where it needs to be. The terminals themselves are precision optical instruments, not commodity boxes.

Acquisition latency is real. A link that takes ten or twenty seconds to come back up after a drop is fine for bulk data routing but uncomfortable for time-sensitive traffic. And unlike an RF crosslink that can broadcast to several listeners at once, one optical terminal serves exactly one peer at a time. Constellations needing dense connectivity have to fly multiple terminals per satellite, which drives up cost and mass.

The terminals themselves have historically been expensive - on the order of one to several million dollars per unit - though high-volume production for Starlink and SDA has driven prices down rapidly. Standards interoperability is another open question. The SDA has published an open Optical Communications Terminal specification aligned with CCSDS standards, but most commercial systems remain proprietary. Two satellites built by different vendors generally cannot link unless they were designed against the same spec, which is a serious obstacle for any future inter-constellation or commercial-defense interoperability.

One last tradeoff worth mentioning: optical links work beautifully between satellites but struggle when one end is on the ground. Cloud cover, turbulence, and atmospheric absorption can collapse a satellite-to-ground optical link in seconds. Pure inter-satellite operation sidesteps this entirely, which is part of why OISLs are an inter-satellite story first and a ground-link story a distant second.

Fun Fact Space Nerds Might Not Know

NASA’s TBIRD demonstration in 2023 hit 200 gigabits per second from a CubeSat in low-Earth orbit, the highest data rate ever achieved from space. The entire laser terminal fit inside a 4U volume - roughly the size of a shoebox - and the receiving ground station was a portable telescope at NASA’s White Sands facility. To put that number in context, 200 Gbps is enough to download the entire Library of Congress text catalog several times over in a single five-minute pass. The same core technology, scaled and ruggedized, is what powers the Tesat and Mynaric terminals flying on Starlink and SDA today.

Looking Forward

SpaceX has flown well over ten thousand Starlink satellites carrying OISLs since the technology debuted on the constellation in 2021, building by far the largest operational laser-mesh network in space. SDA’s Tranche 0 is on orbit, Tranche 1 is launching now, and Tranche 2 has already been procured - all of it built around the Transport Layer’s optical mesh. GPS III SV10’s OISL demo is the first time the technology has flown on a GPS satellite, and it points toward a future where navigation, communications, and missile-tracking constellations share data laterally instead of bouncing every byte through the ground.

The unresolved fights are about standards. SDA is pushing hard for the open OCT spec to become the lingua franca for U.S. government and commercial systems alike. Several large commercial operators prefer their proprietary stacks. Whoever wins shapes how interoperable the space-based internet of the next decade actually is. Beyond that, researchers are already extending OISL technology into quantum key distribution - using single-photon properties of laser links to exchange encryption keys with provable security guarantees. The first space-based QKD demonstrations have already flown, and they ride on the same fundamental optical hardware.

For KeepTrack users, OISLs are mostly invisible by design - they leave no obvious orbital signature - but they shape what you see in subtle ways. Many active satellites in modern proliferated constellations no longer have a tight relationship with any specific ground station. A Starlink in the middle of the Pacific is not silent; its data is routing through a chain of laser-linked peers until it reaches a satellite that does have a ground station in view. When you watch a constellation pass overhead in KeepTrack and wonder how a satellite over open ocean is serving customers in Europe, the answer is almost always a beam of infrared light traveling between two of those tiny moving dots.