Q-RELAY
This memory-assisted quantum repeater for standard fibre networks is designed around a control layer that Quantolio owns: the real-time synchronization and adaptive control that coordinate its quantum memories and Bell-state measurements.
These values define the Phase I design envelope and targets. No Q-RELAY results have been measured yet.
01 Investment Highlights
Q-RELAY targets the integrated repeater controller, a role no commercial product fills yet.
Quantum networks need repeaters to reach long distances over standard fibre. Quantolio is developing the control platform for those repeaters with federal grant funding, and it keeps ownership of the resulting intellectual property.
No integrated commercial repeater controller exists yet
Laboratories have demonstrated the components, but no commercial product yet runs memories, measurements, scheduling and recovery as one repeater node.
Quantolio owns the control-layer IP
Quantolio buys detectors and plans to source memories and photon-pair sources through laboratory partners. It builds the control software and firmware and owns all foreground intellectual property (IP).
The platform is not tied to one memory material
A hardware-abstraction layer lets different memory materials share the same device and network interfaces.
One control layer serves two federal programmes
Q-RELAY uses the same field-programmable gate array (FPGA) control layer as SENTINELLE-Q, which Innovative Solutions Canada selected for federal testing in September 2026.
Phase I is approved for non-dilutive federal grant funding
Phase I is approved under Innovative Solutions Canada. If Quantolio is invited, Phase II would fund a physical two-link prototype over up to 24 months.
Every milestone has a measurable success criterion
Phase I has six milestones over 26 weeks. If Q-RELAY does not exceed direct transmission, it would still deliver a quantified threshold roadmap.
02 The Market Problem
Long-distance quantum networks need repeaters, because the rate of a direct fibre link falls with distance.
Fibre loss limits direct transmission
Standard fibre loses 0.2 to 0.3 dB per kilometre near 1550 nm. Without a repeater, the entanglement rate cannot exceed the repeaterless rate-loss bound.
Stored states must survive until the adjacent link succeeds
A repeater holds one half of each link in memory until the neighbouring link succeeds. Memory lifetime and efficiency are therefore among the dominant feasibility risks.
Entanglement swapping requires indistinguishable photons
Each Bell-state measurement depends on interference between two photons, so their frequency, timing and polarization must match. Q-RELAY targets a visibility of 90% or more.
03 The Solution
Quantolio’s contribution is the control platform designed to run the whole repeater node.
Q-RELAY is designed to connect two fibre links through a central repeater node. Quantolio builds the control and network software, and a hardware-abstraction layer keeps it independent of any single memory material.
Current status
A comparative engineering study selected the two-link architecture and its Phase I test envelope. No integrated laboratory result is claimed yet.
Network software
The network layer handles device interfaces, scheduling, swapping coordination, routing and telemetry, and it will track every stored mode by age, fidelity and expiry.
FPGA node controller
The FPGA controller will handle deterministic timing, both memory banks, 16 to 32 temporal modes, time tagging, alarms and recovery.
Quantum hardware
The hardware layer consists of rare-earth quantum memories, narrowband photon-pair sources, telecom interfaces and superconducting nanowire single-photon detectors (SNSPDs).
04 How It Works
The repeater protocol runs in five steps, from photon-pair generation to end-to-end entanglement.
The FPGA controller will run this sequence for every temporal mode. Phase I must run it over at least 1,000 consecutive hardware-in-the-loop (HIL) cycles without manual intervention.
The schematic is not to scale in distance or time.
What happens
In the reference configuration, each source produces a photon pair by cavity-enhanced spontaneous parametric down-conversion (SPDC) in periodically poled lithium niobate (PPLN): a 606 nm photon for the memory and a native telecom-band partner for the fibre.
What happens
The memory photon is stored locally in a rare-earth atomic frequency comb (AFC) memory, in one of 16 to 32 temporal modes. Its partner travels through standard fibre toward a midpoint station.
What happens
Photons from opposite ends interfere at a 50:50 fibre beam splitter. Superconducting nanowire single-photon detectors (SNSPDs), FPGA time tagging and coincidence logic herald the entanglement and flag which stored mode succeeded.
What happens
The controller waits until the repeater's left and right memories both hold valid states, then selects which stored modes to retrieve.
What happens
Synchronized retrieval followed by a local Bell-state measurement swaps the entanglement out to the two end nodes.
05 Temporal Multiplexing
Multiplexing gives each link attempt 16 to 32 chances to succeed.
Fibre loss means a single photon pair rarely succeeds over a long link. Q-RELAY sends photon pairs in 16 to 32 time bins per attempt, stores each in its own memory mode and keeps whichever one is heralded, so the success rate per attempt rises with the number of modes.
Here p is the success probability of one mode and N is the number of modes.
The animation uses an arbitrary success probability per mode and does not show Q-RELAY performance.
- Phase IIThe Phase II prototype would characterize operation with 16 to 32 modes in physical memories.
06 First Differentiator
Fidelity-aware scheduling decides which stored states to use, keep or discard.
Today’s baseline is fixed first-in, first-out (FIFO) control. Q-RELAY will score every stored mode instead, and this scheduling logic will form part of Quantolio’s foreground IP.
The animation shows how the controller tracks modes and does not represent measured data.
Each stored mode is scored before it is used
- Age
- Estimated fidelity
- Retrieval probability
- Adjacent-link readiness
- Expected BSM success
higher fidelity-adjusted entanglement rate
fewer memory modes at equal rate and fidelity
lower discard rate
higher delivered fidelity at a fixed rate
Improvements will be claimed only under identical test conditions, with both positive and negative operating regions reported.
07 Second Differentiator
Stabilization and fault recovery are designed into the node controller.
The controller will detect timing, detector, memory and link faults, suspend invalid attempts, recalibrate and restore operation without an operator. Today’s expert-operated setups cannot run unattended for hours.
- Timing
- Detector health
- Memory state
- Link loss
- Frequency, phase and polarization
The run will measure uptime, event loss, detection rate, false alarms, recovery time and manual interventions.
The animation illustrates the test procedure. Fault markers are illustrative and do not represent the test schedule.
detection of supported injected faults
false alarms
to initiate recovery
to complete a supported recovery
fewer manual interventions than the open-loop baseline
08 The Benchmark
Q-RELAY will be benchmarked against the repeaterless limit of direct transmission.
Phase I will compare Q-RELAY with the repeaterless rate-loss bound, known as the PLOB bound, which is the upper limit for any direct fibre link.
Q-RELAY must exceed this bound in at least one physically credible operating region. Its own curve is a Phase I output and is therefore not drawn.
Phase I will report where Q-RELAY exceeds direct transmission, with every input traced to a measurement, a published result or a component specification.
Phase I will deliver a quantified roadmap of the minimum memory, source, detector and multiplexing performance a repeater needs, and it will claim no unsupported advantage.
09 Phase I Programme
Phase I is a 26-week programme with six milestones, each tied to a measurable success criterion.
The milestones overlap, and the digital twin runs alongside the feasibility and controller work.
Deliverable
Confirm the memory and source choices and freeze the architecture, baselines and acceptance tests.
Success criterion
Every outcome is mapped to a test, and one memory and one source and channel architecture are selected.
Deliverable
Build the digital twin and the repeaterless benchmark, with sensitivity and uncertainty analysis.
Success criterion
A validated model gives rate and fidelity against distance for direct transmission, fixed scheduling and adaptive Q-RELAY.
Deliverable
Model memory storage, efficiency and fidelity across 16 to 32 modes, complete the source and link-budget studies, and run optical measurements.
Success criterion
The preferred memory and channel interface are selected, and the minimum component thresholds for a crossover are documented.
Deliverable
Implement deterministic timing, the dual-memory state machines, 16 to 32 temporal modes, telemetry, adaptive scheduling and recovery on the FPGA.
Success criterion
The complete two-link sequence executes deterministically in hardware-in-the-loop tests.
Deliverable
Connect the digital twin, the FPGA and the emulators over representative fibre impairments, compare scheduling policies and run unattended tests with injected faults.
Success criterion
At least one scheduling improvement of 25% or more (or 20% or more in fidelity) is shown, and the eight-hour unattended run meets the fault-detection and recovery thresholds.
Deliverable
Compare performance with the repeaterless benchmark and complete the prototype, procurement, IP and commercialization plans.
Success criterion
Phase I ends with a credible crossover region or a quantified threshold roadmap, plus a complete Phase II prototype specification.
10 Phase II, If Invited
Phase II would take Q-RELAY from feasibility to a physical two-link prototype.
The prototype would contain two fibre links and a central repeater node with physical rare-earth memories, photon-pair sources, detectors, FPGA controllers and network software.
Team
Phase II would need about 10 to 12 full-time-equivalent staff, with Quantolio performing at least half of the research and development.
Barriers and mitigation
The identified barriers include memory performance, cryogenic reliability, photon indistinguishability and long-lead components. Staged subsystem gates, alternate memory and wavelength options, early procurement and milestone-gated spending address them.
Next Steps
Discuss Q-RELAY with the team
Talk to the team about the investment case, quantum networks, distributed sensing, or laboratory collaboration on memories, sources and optical testing.
Investor briefing
Review the investment case, the IP position and the Phase I milestones with the team.
Research collaboration
Partner with the team on laboratory work involving memories, photon-pair sources and optical tests.
Phase II scoping
Discuss the scope of the two-link prototype, its subsystem gates and its procurement.