A team of physicists has shown that quantum-encrypted keys can travel 75 miles through the same fiber-optic cable already carrying ordinary internet traffic, removing one of the biggest practical barriers to quantum-secured communications. Their results, accepted for publication in Physical Review Letters and detailed in a public preprint, demonstrate that a quantum signal can survive sharing a glass strand with high-power conventional data channels over 120 kilometers of standard telecom fiber. The preprint, posted in February 2025 and accepted by PRL in the months that followed, describes work that until recently had no parallel: most long-distance quantum key distribution experiments required a dedicated, otherwise empty line. This one did not.
That distinction matters because no telecom carrier wants to lay entirely new cable just for quantum security. If quantum encryption can ride on infrastructure that already exists, the path from laboratory proof to real-world deployment gets dramatically shorter.
Why this experiment is different
Quantum key distribution, or QKD, uses the laws of quantum physics to generate encryption keys between two parties. Any attempt to intercept the key disturbs the quantum states involved, alerting both sides to the intrusion. In principle, this makes the key exchange provably secure in a way that mathematical encryption alone cannot guarantee.
The technique used here is called continuous-variable QKD (CV-QKD). Instead of encoding information on individual photons, which demands specialized and expensive single-photon detectors, CV-QKD encodes it in the amplitude and phase of light waves. That is a significant advantage: it means the quantum signal can be picked up by the same type of detector hardware telecom companies already use, cutting one of the cost barriers to adoption.
Previous experiments had pushed QKD to impressive distances on isolated fiber. But those trials typically ran on “dark fiber,” a cable carrying nothing but the quantum channel. In a real metropolitan network, that same cable is packed with dozens or even hundreds of wavelength-division-multiplexed data channels, each one a potential source of optical noise that could drown out the delicate quantum signal.
The new experiment confronts that noise problem directly. By carefully managing wavelength assignments, filtering, and power levels, the researchers showed that CV-QKD can coexist with classical data streams on a single 120 km span and still produce secret keys at rates sufficient for real encryption after error correction and privacy amplification.
Supporting work from NIST and other groups
This result does not stand alone. A NIST study previously demonstrated that a different flavor of QKD, called measurement-device-independent QKD, could coexist with multiple 10 Gbps classical channels on the same fiber and still recover secure keys. That work focused on the coexistence question rather than raw distance, and it reinforces the broader point that quantum and classical signals can share infrastructure without catastrophic interference. Separately, experiments published in Scientific Reports have shown high-speed encrypted data transport integrated with QKD on shared fiber, bridging the gap between generating a quantum key and actually using it inside a live encryption pipeline.
Together, these results establish a consistent pattern: quantum security does not require telecom operators to rip out their infrastructure. It can be layered on top of what already exists, provided the optical engineering is done carefully.
The threat that gives this urgency
The research is driven in part by a scenario intelligence agencies call “harvest now, decrypt later.” Adversaries can record encrypted communications today, store them, and wait for a sufficiently powerful quantum computer to crack the encryption years from now. Sensitive diplomatic cables, financial transactions, and medical records intercepted in 2026 could become readable in 2035 or 2040 if large-scale quantum computers arrive on that timeline.
No quantum computer currently exists that can break the RSA or elliptic-curve cryptography protecting most internet traffic. But the threat is taken seriously enough that NIST finalized its first set of post-quantum cryptography standards in August 2024 (FIPS 203, 204, and 205), giving software developers new algorithms designed to resist quantum attacks. QKD and post-quantum cryptography address the problem from different angles: PQC replaces vulnerable math with harder math, while QKD replaces math-based key exchange with physics-based key exchange. Many security planners expect both approaches to be deployed together in high-value networks.
What still has to go right
Several hard problems sit between this laboratory success and a product a telecom carrier could buy.
Channel density. Real metropolitan fiber networks carry dozens to hundreds of wavelength channels simultaneously. Whether the quantum signal remains viable as that channel count scales up is the central engineering question the field still faces. The 120 km experiment demonstrated coexistence with classical traffic, but the full picture of how many simultaneous channels were present and at what power levels requires close reading of the primary paper.
Key rates. QKD faces a fundamental tradeoff: longer distances mean weaker quantum signals, which means fewer secure key bits per second. At 120 km, the system generates keys, but whether the rate is fast enough to refresh encryption for high-bandwidth metropolitan traffic has not been independently confirmed. In most realistic deployments, QKD keys would periodically refresh symmetric encryption keys rather than encrypt every bit directly, but even that approach has minimum rate requirements.
Real-world fiber conditions. A 120 km spool in a controlled lab is not the same as 120 km of buried urban cable with its splices, bends, aging connectors, temperature swings, and variable signal loss. Quantum states are fragile, and every imperfection in the link chips away at the signal.
Network operations. Telecom operators rely on dynamic routing, automatic power balancing, and frequent reconfiguration of wavelength assignments. A QKD system that breaks every time a new wavelength channel is added or an amplifier is retuned would be impractical for production networks. Current experiments generally assume stable, well-characterized conditions.
Side-channel vulnerabilities. QKD’s security guarantees are rooted in physics, but the hardware that implements it is not immune to engineering flaws. Detector blinding attacks and other side-channel exploits have been demonstrated against earlier QKD systems. Calling any real-world implementation “unbreakable” requires confidence not just in the protocol but in every component of the physical device.
Where quantum-secured networks stand in mid-2026
Several countries are already operating or building QKD networks. China’s Beijing-Shanghai backbone, anchored by the Micius satellite for intercity links, has been running since 2017. The European Union’s EuroQCI initiative aims to connect all 27 member states with quantum communication infrastructure. South Korea and Japan have active pilot programs. Most of these networks, however, rely on dedicated fiber or trusted relay nodes, both of which add cost and complexity.
The 120 km shared-fiber result matters because it attacks the cost problem at its root. If quantum keys can share existing cables, carriers do not need to lease or light separate fiber strands for every quantum-secured link. That changes the economics from “expensive niche tool for governments and banks” toward something that could, eventually, protect a much wider slice of internet traffic.
The remaining questions are not small. Scaling channel density, sustaining useful key rates over imperfect real-world routes, integrating QKD into live network management, and hardening devices against side-channel attacks will each take years of engineering. But the coexistence barrier was among the most commercially relevant, and as of mid-2026, it looks increasingly like a solved problem rather than an open one.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
