Maturity — TRL across platforms
A procurement-ready quantum platform is one that has graduated from basic research through prototype demonstration to operational deployment. NASA's Technology Readiness Level (TRL) scale measures this graduation on a 1–9 ladder. Purohit et al. 2023 adapt this to the quantum context as QTRL, with quantum-specific milestones at each level and an S-curve framing of how fast progress happens at each step Purohit et al. 2023 .
Applied to today's quantum-network hardware, the picture is uneven. Some components (commercial QKD, certain memory platforms) sit in the 5–7 range — field-deployable. Others (microwave-to-optical transduction, fault-tolerant logical qubits) sit at 2–4, still pre-prototype. Meddeb 2025 Table 5 gives concrete ratings for the per-platform memory landscape Meddeb 2025 .
The QTRL framework
The QTRL ladder mirrors NASA TRL semantics but anchors each level in a quantum-specific milestone Purohit et al. 2023 :
- QTRL 1–2 — basic principles observed; concept formulated.
- QTRL 3 — analytical or experimental proof-of-concept.
- QTRL 4 — component validated in lab.
- QTRL 5 — component validated in relevant environment.
- QTRL 6 — system prototype demonstrated in relevant environment.
- QTRL 7 — system prototype demonstrated in operational environment.
- QTRL 8 — actual system completed and qualified.
- QTRL 9 — actual system proven in operational environment.
Purohit's framing adds an S-curve observation: progress is slow at QTRL 1–3 (fundamental physics still under exploration), accelerates through 4–6 (the engineering payoff phase), then slows again at 7–9 (operational hardening dominates). A platform that has been at the same QTRL for several years without a clear blocking issue is usually mid-S-curve and likely to advance soon; one stuck at QTRL 3 for a decade is more likely to be limited by physics rather than engineering.
Per-platform memory TRLs (Meddeb 2025 Table 5)
Meddeb's survey provides the cleanest published per-platform TRL benchmark for quantum-memory hardware. The table below reproduces the ratings (TRL columns are Meddeb's; the surrounding context comes from the cited paper) Meddeb 2025 .
| Platform | Best storage time | QTRL band | Reference demo |
|---|---|---|---|
| Atomic ensembles (cold atoms, rare-earth crystals) | ~1 hour | 7–8 | Ytterbium-171, Eu:YSO crystals |
| Trapped ions | ~10 s (network-grade) | 6–7 | Multi-node trapped-ion demos |
| Neutral atoms | ~seconds | 5–6 | Optical-tweezer arrays |
| NV / SiV centres (diamond) | ~100 µs (electron); s (nuclear) | 5–6 | Delft / Harvard multi-node demos |
| Quantum dots | ~ms–days (depending on encoding) | 5–6 | Indistinguishable-photon emitters |
| Photonic delay-line memories | ~µs | 4–5 | Optical-fibre delay loops |
Capability-aware ratings
A single TRL per platform misses the structure of the field. Most platforms are good at one role and worse at others:
- Trapped ions are mature for compute (QTRL ~7 for small-scale machines) but slow for networking — the photonic interface for remote entanglement runs at lower rates than transmons or atoms.
- Superconducting transmons are mature for compute (QTRL ~7 for noisy intermediate-scale machines) but lack a native optical interface; the microwave-to-optical transduction problem sits at QTRL ~3.
- NV centres sit at QTRL ~6 for matter-photon entanglement (the link-layer interface that memory-based repeaters need) but lower for high-fidelity computation, since gate fidelities lag the leading compute platforms.
- Photonic dual-rail is well-suited to communication (QTRL ~6 for QKD endpoints) but room-scale fault-tolerant photonic compute remains QTRL ~3.
Where the bottlenecks are
Across the field, four areas sit at low TRL and dominate the gating roadmap to deployable quantum networking:
- Microwave-to-optical transduction (QTRL ~3) — until coherent efficiency rises above ~1 %, superconducting quantum computers can't distribute entanglement off-chip without a heterogeneous matter-platform bridge.
- Multi-mode quantum memory (QTRL ~5) — single-mode storage is mature; storing many qubits in parallel (needed for multiplexed repeater designs) is still pre-engineering.
- High-rate indistinguishable single-photon sources (QTRL ~5–6) — heralded photon sources at the rates and indistinguishabilities all-photonic repeaters need are not yet routine.
- Fault-tolerant logical qubits at network scale (QTRL ~3) — third-generation quantum repeaters need this; current prototypes do at most a handful of logical operations.
Forward references: the transduction, memories, repeaters, and companies subjects unpack these bottlenecks and the vendors working on them Kumar et al. 2025 .