The Core Infrastructure Difference
Fiber-based QKD and satellite QKD both deliver information-theoretic key material, but they start from opposite infrastructure assumptions. Fiber QKD requires a continuous, unbroken optical fiber path between the transmitter and receiver. Satellite QKD uses a free-space optical link between an orbiting satellite and ground stations, with the key material relayed through space rather than through glass.
This difference is not cosmetic. It determines where each approach is deployable, what its cost structure looks like, and which operational risks dominate. For most infrastructure operators considering QKD deployment, the choice between fiber and satellite comes down less to physics preferences and more to what the existing network topology makes physically possible.
What Fiber QKD Requires
Fiber QKD requires dedicated dark fiber between every pair of nodes that need to exchange quantum keys directly. This is not an off-the-shelf capacity arrangement. Commercial fiber is shared, multiplexed, and often traverses optical amplification stages that destroy the quantum states required for BB84 or similar protocols. QKD requires either purpose-laid fiber or dark fiber that has been verified clean of optical tap points and dispersion compensation elements that damage single-photon coherence.
For a bank wanting to establish QKD between its primary data center and its disaster recovery site in the same metropolitan area, this is a manageable requirement. Both sites are probably already connected by multiple fiber paths, and procuring a dark fiber lease for a dedicated QKD channel is a known operational task, even if it takes months to complete.
The trusted repeater requirement is a more fundamental constraint. QKD over fiber is limited by photon loss to approximately 100 to 150 kilometers without intermediate nodes, depending on fiber quality. Beyond that distance, the key material must be regenerated at trusted relay nodes: physical locations where the quantum channel terminates, the key is extracted from quantum to classical, re-encrypted, and passed to the next quantum segment. These relay nodes are trusted in the cryptographic sense, meaning their compromise breaks the security guarantee for the full path. Building a national network with fiber QKD means building and securing trusted relay infrastructure across the full coverage area.
What Satellite QKD Removes
Satellite QKD removes both the dedicated dark fiber requirement and the trusted relay node requirement, for the nodes it can reach.
In free-space QKD using a satellite as the distribution point, a single satellite in low or medium Earth orbit can exchange key material with any ground station within its field of view during a pass. No fiber connection between the ground stations is required. Two cities on opposite sides of a country, each with a QKD ground station, can receive correlated key material from the same satellite pass without any physical connection between them. The satellite acts as an untrusted relay: because it distributes entangled photon pairs to both endpoints simultaneously (in E91-based approaches), or because it generates a key that is never stored intact on the satellite (in other architectures), the satellite itself does not need to be trusted in the classical sense.
This changes the deployment calculus significantly for multi-site or geographically dispersed networks. A national bank with branches in 15 cities cannot economically run dedicated dark fiber between all of them. It can, in principle, operate a ground station at its primary locations and share satellite passes for key material delivery.
The Last Mile: What Satellite QKD Leaves Unsolved
Satellite QKD does not solve the last mile problem. Getting key material from the ground station to the network equipment that needs it at each location still requires a local key management infrastructure, typically a key management server connected to the ground station via a secured local network segment.
The coverage window is also a real operational constraint. A single satellite in low Earth orbit typically has a ground pass window of 10 to 15 minutes over a given city. During atmospheric conditions that reduce optical link quality (cloud cover, heavy precipitation, aerosol loading), the quantum bit error rate may exceed the threshold at which secure key distillation is possible. Key buffers accumulated during good-weather passes provide continuity, but sizing those buffers requires knowing the expected demand and the statistical distribution of weather disruption at each site.
This is where the AI scheduling layer becomes practically necessary rather than theoretically nice. We have observed in our Bengaluru pilot that naive static scheduling of satellite passes wastes a significant fraction of available key generation capacity because it does not account for ground-station availability, atmospheric conditions, and the current state of key buffers at each node. Adaptive scheduling that uses weather forecast data and orbital geometry to prioritize passes where the link quality and buffer state will benefit most from key generation can increase effective key delivery by 30 to 40 percent compared to round-robin scheduling.
Metropolitan vs. Regional vs. National Scale
The practical fit between fiber and satellite QKD shifts significantly by scale.
At metropolitan scale, two to five sites within a single city, fiber QKD is often the more immediately practical option. Dark fiber exists in most urban business districts. The distance is within single-segment QKD range. The operational model of a dedicated, always-on quantum channel matches the reliability expectations of financial or data center applications.
At regional scale, three to ten sites spanning hundreds of kilometers, the trusted relay node requirement for fiber QKD starts to dominate costs and operational risk. Each relay node is a security liability that must be physically secured and included in the threat model. Satellite distribution for the inter-city key material, with fiber QKD or secure classical channels handling intra-city distribution, becomes an architecturally cleaner approach.
At national scale, satellite QKD is the only practically deployable approach for broad geographic coverage without building extensive trusted relay infrastructure. This is why the early adopter countries for national QKD networks have converged on satellite as the backbone, with metro-scale fiber QKD for high-density urban nodes.
Hybrid Architectures in Practice
The most thoughtful deployments we have seen in planning discussions are not choosing between fiber and satellite QKD: they are using both, with satellite handling the long-distance metropolitan-to-metropolitan links and fiber handling the high-bandwidth intra-site connections where distance is not a constraint.
This is not an observation that favors one technology over the other. It reflects the actual topology of real networks, which have different distance and bandwidth requirements at different tiers. The key management architecture needs to be able to ingest key material from both sources and maintain appropriate accounting of which key material came from which source, since the trust properties and availability patterns differ.
We are not saying fiber QKD is a worse approach than satellite. It is a more immediately deployable approach for single-city applications where dark fiber is accessible. Satellite distribution is the more practical approach for wide-area coverage and for operators who cannot procure dedicated fiber paths between their sites. The decision is primarily architectural and operational, not a judgment about physics.
