Quantum communications uses single photons and entanglement to move keys and quantum states between distant parties with security guaranteed by physics rather than computational hardness; the point-to-point hardware is now working at telecom wavelengths on installed fibre, but the memories and repeaters needed for actual networks remain proposals.
Summary
Quantum communications covers two related things. The first is quantum key distribution (QKD), where a sender encodes bits on single photons or weak coherent states and any eavesdropper’s measurement necessarily disturbs them, so the legitimate parties can bound the leaked information and distil a key whose security rests on quantum mechanics rather than on an unproven computational assumption. Protocol families differ in what they assume: prepare-and-measure schemes such as BB84 and SARG04, entanglement-based schemes such as BBM92, and continuous-variable schemes that encode on the quadratures of light and can be built from standard telecom transmitters and homodyne receivers. The second, more ambitious thing is a quantum network that distributes entanglement itself, as a resource for distributed quantum computing, distributed sensing and communication primitives that have no classical equivalent.
The parameters that decide the field are all physical-layer. Photon loss in fibre is exponential in distance and cannot be amplified away without destroying the quantum state, so key rate falls steeply with range: a recent hybrid-integrated receiver achieved over 2 Mbps at short distance and 15 kbps over 100 km of fibre using BB84. Detector quality sets the noise floor, and detector technology forces a hard trade: superconducting nanowire detectors have been integrated onto receiver chips but need cryogenics, while gated InGaAs/InP single-photon avalanche diodes work without cooling and can now be built into GHz-gated arrays with negligible inter-pixel crosstalk. For free-space and infrared operation the cryogenic problem is worse still, with existing mid- and long-wave infrared single-photon detectors needing 0.08 to 0.9 K.
Beyond a single link, distance requires quantum repeaters, and repeaters require both high-rate heralded entanglement sources and photon memories. Both are at the design and proof-of-principle stage in the sources here: a zero-added-loss multiplexing source design has been proposed for time-bin entanglement with experimental verification only of the compatibility of time-bin pulses and spectral shearing, and single-photon storage via Brillouin interactions in nanowires is a theoretical configuration, with the acoustic-wave lifetime identified as the limiting factor.
The most consequential deployment result is coexistence. Polarisation-encoded entanglement has been distributed in the O-band over an installed 24.4 km fibre while the same fibre carried a fully loaded state-of-the-art C-band classical line system with two 800 Gbps channels and the rest of the C-band filled with amplified spontaneous emission, plus a picosecond-precision L-band synchronisation signal, with Bell state fidelity preserved by careful wavelength allocation and narrow filtering against spontaneous Raman scattering. That determines whether quantum channels need dark fibre or can ride existing plant, which is the single largest cost question in the field.
Evidence base
- Hybrid receivers combining GHz-gated InGaAs/InP SPAD arrays with low-loss silica waveguide chips ran BB84 at over 2 Mbps at short distance and 15 kbps over 100 km of fibre, without cryogenic cooling, reported 5 September 2025.
- O-band polarisation-encoded entanglement was distributed over an installed 24.4 km fibre while the same fibre carried a fully loaded C-band classical line system with two 800 Gbps channels plus ASE fill and an L-band picosecond synchronisation signal, with Bell state fidelity preserved (updated 21 July 2026).
- A finite-size security analysis for discrete-modulated continuous-variable QKD against general attacks removed bounded-dimension assumptions and reported positive key rates at block sizes of order 10^8 with a trusted detector model, 19 June 2026.
- A quantum ring all-reduce primitive using pre-shared entanglement and superdense coding reduces per-link online communication by a provably optimal factor of two and gives composable epsilon-secure aggregation at 2x overhead in GHZ copies, 19 June 2026.
- Repeater-grade entanglement sources remain at the design stage: a ZALM time-bin design was proposed with experimental verification only that time-bin pulses and spectral shearing are compatible (no appreciable phase shift), 5 May 2026.
- Existing mid- and long-wave infrared single-photon detectors for free-space quantum communications require 0.08 to 0.9 K; a ferromagnet/superconductor bilayer approach proposes raising this to 3.75 K, 17 June 2026.
- NSF funded a hybrid free-space and fibre quantum network programme with a satellite entanglement backbone at $368,000 (Yale) and $317,680 (UMass Amherst) on 2 August 2026, and $473,811 for quantum detection and estimation theory on 4 August 2026.
Open questions
- Can any photon memory reach a storage lifetime and efficiency sufficient for a two-link repeater chain, given that the Brillouin nanowire scheme is limited by acoustic-wave lifetime and remains theoretical?
- Does the O-band coexistence result at 24.4 km scale to full metro spans and multi-node topologies, or does spontaneous Raman noise from fully loaded C-band systems dominate beyond a certain length?
- What is the cost per secure bit of a hybrid-integrated SPAD receiver relative to classical alternatives, given that none of these sources contains cost data?
- Does the proposed ZALM time-bin source, once built, actually deliver quasi-deterministic heralded entanglement at rates useful for repeaters, rather than only component-level compatibility?
Assessment drafted 2026-08-31 from up to 15 KB sources using the technology-scorecard framework; scores are a draft read pending review.