Quantum Networking Ecosystem Analysis

In their simplest description, quantum networks enable quantum information to be sent or shared between three or more nodes. Quantum networking is therefore an umbrella term that encompasses technologies that address drastically different functionalities, length scales, use-cases, customers, and markets. However, the length of these optical links, the geometry of the network architecture, how light from each node enters the network, and the role of the network itself, are all design choices that depend on the application each company focuses on. Nu Quantum is currently the only company working exclusively on building quantum networks for distributed quantum computing.

‍

Broadly speaking, the applications of photonic quantum networking include:

Quantum communication: long-range communications (eg. satellite), quantum internet, cybersecurity (including quantum key distribution);
Quantum sensing: networks of quantum sensors may present unique advantages for metrology;
Distributed quantum computing (dQC): networked quantum processors will enable multi-core quantum computing.

The distributed quantum computing use-case requires the development of hardware solutions that address the many functions underpinning a network — interfacing with qubits, photonic switching technology to enable entanglement distribution, and network orchestration and control — as well as a blueprint for how these must come together into an architecture that will allow quantum algorithms to run efficiently with high rate and fidelity across the network.

There are only ~10 companies worldwide focusing on quantum networking solutions for the above applications, most of whom develop a subset of the above components. Nu Quantum is the longest-standing in the sector, with the largest team and widest breadth of technology. We are also the only company working exclusively on developing full-stack quantum networking hardware solutions, backed by an architecture compatible with distributed QEC — which come together to form the networking layer we call the Entanglement Fabric — to enable distributed quantum computing.

A number of networking companies focus on developing components called quantum memories, which are required in some quantum networking protocols to ensure that quantum information is preserved as it is transmitted between different network nodes. To date, these companies have not announced activity in other layers of the networking stack (QPI, QNU, architecture). Each memory company uses a different material platform (neutral atoms, solid-state, or trapped ions) as their memory, such that integration with dissimilar qubit systems may present challenges and instead may be most compatible with specific computing modalities. As Nu Quantum’s distributed computing architecture does not require a quantum memory (deep-dive below), we believe that the interests of these companies are complimentary to our vision at this stage.

Competitive Analysis*

{{competitive-analysis="/rich-text-content/tables/quantum-networking-ecosystem-analysis"}}

*The information provided in this section is correct to the best of our knowledge using publicly available information of company structures and roadmaps as of 14 October 2024.

A summary of each company and their current activity is provided below, grouped by the types of solutions that they provide.

Quantum networking companies

Aliro Quantum (software) 🇺🇸

Aliro is a software company based in Boston (USA), specialising in building advanced secure networks through a proprietary operating system. They spun-out of Harvard University in 2019, founded by Dr Prineha Narang (CTO) and Dr Michael Cubeddu, and is led by CEO Jim Ricotta. Their Series A fundraise was backed by Leaders Fund, Accenture, and Cisco Investments.
Their networking software solutions target applications in post-quantum cryptography, quantum key distribution and quantum-secure communication. Currently, they provide a network orchestration simulator to help companies emulate and evaluate networks. They list distributed quantum computing as a possible use-cases of their software solutions, but do not have any public partnerships with quantum computing companies, or quantum networking hardware providers. They do not develop networking hardware solutions.

Bohr Quantum 🇺🇸

Bohr Quantum is a spinout of Caltech in 2021, formed by Paul Dabbar, former Under Secretary for Science at the US Department of Energy and Conner Prochaska, former Chief Commercialisation Officer for the US Department of Energy. They focus on networking technologies for emerging quantum internet applications, but do not have any publicly available information on their R&D activity, products, or other employees.

Cisco Research 🇺🇸

Cisco Research, the R&D division of global networking giant Cisco, has been interested in quantum for several years, including as a commercial partner on two projects led by Nu Quantum and funded by the UK government: Medusa (2022-24) and Lyra (2023-ongoing). Cisco Research’s quantum research team was formally established in early 2024 in Santa Monica (California, USA) and is led by Dr Reza Nejabati, former Professor and Research Group Leader in Quantum Networking at Bristol University (UK). Their research interests (presented at Cisco Live 2024) are varied: quantum cryptography, secure communications, global clock synchronisation, distributed QC, blind QC, distributed sensing, and quantum money.
Cisco’s quantum team may develop ‘competing’ technology to Nu Quantum, particularly around the quantum networking control and orchestration or photonic layer. Their work within the distributed QC use-case does not currently address development of qubit-photon interfaces or quantum processors. We believe that quantum systems based on networking will require dedicated networking solutions that take into account the nature of the subsystems which they control.
Therefore, we believe that Cisco and Nu Quantum’s commercial interests are complimentary at this stage and this is reflected in Cisco Research’s relationship with us as lead-customer/end-user of our ongoing Lyra project to deliver a world-first quantum networking unit prototype.

Qunnect 🇺🇸

Qunnect is based in New York City and was spun-out from Stony Brook University in 2019, and is led by Dr Mehdi Namazi CSO, Mael Flament CTO, Noel Goddard CEO. The company focuses on developing hardware modules that enable the realisation of long-range quantum networks, with particular emphasis on the quantum internet. Recently, they demonstrated the generation, distribution, and preservation of entangled photons over a 34 km existing commercial fibre network in New York City.
Their products include a quantum entanglement light source for high interfacing efficiency, a quantum memory component with room-temperature operation, wavelength references for laser stabilisation, and automated polarisation compensating devices. Their solutions are currently aimed at enabling and maintaining long-range fibre-based network connectivity, and their ongoing research aims to produce ways to monitor the quality of distributed entangled photons.
Although they list distributed quantum computing as a potential use-case of their various network solutions, they do not focus on delivering network architectures for distributed computing. We believe that Qunnect’s activity and expertise in maintaining the quality of long-range photonic links is complimentary to the scientific and commercial interests of Nu Quantum.

Qubit-photon interface (QPI) companies

NanoQT 🇯🇵

NanoQT is Japan’s first quantum computing hardware company and was founded in 2022 to commercialise the nanofibre-based reserach developed by Professor Takao Aoki and his team at Waseda University. Their core technology is a nanosized fiber resonator for neutral atom qubits, which in principle could 10,000 atoms in a single optical microcavity. The nanofibre cavity in which the neutral atoms are trapped natively allows light from the qubits to be extracted. They propose that interconnecting multiple nanofiber cavities together would enable the formation of a distributed computing network.
This approach is a fibre-based analog of the qubit-photon interfaces that Nu Quantum is developing, which are instead based on open optical microcavities (link to NQ’s QPI explainer). In our approach the qubits do not live inside a fibre — our QPI instead allows us to direct light from the qubits into the fibre (we believe this has advantages for stability, as qubits are not directly exposed to the vibrations and noise of the outside world via the fibre). NanoQT’s technology is exciting and may be particularly suitable for neutral atoms once qubit integration is demonstrated, but is not being developed for other qubit modalities.

‍

Transduction companies (MW to optical conversion)

Miraex 🇨🇭

Miraex was spun out of EPFL (Switzerland) in 2019, although the original co-founders no longer serve on the executive suite. Miraex develop microwave-to-optical transducers for applications in sensing and computing, enabling microwave signals to be converted to C-band telecom wavelengths. They are part of the IBM Q Network for quantum computing. Miraex has also received 3.2M USD project to scale their quantum sensing technology, and was selected to support NASA’s remote sensing programme.

QPhox 🇳🇱

QPhox was spun out of the University of Delft (Netherlands) in 2021, and is led by Dr Rob Stockill (CTO) and Prof Simon Groeblacher (CEO), who also heads an academic research group at the university. They develop low-loss and high-fidelity transducers for microwave to optical frequency conversion, with technology based on mechanical resonators triggered via piezo-electric and optomechanical effects. These components are crucial for interfacing superconducting qubits (which emit photons at microwave frequencies) with photonic networks (which operate in the optical/telecom frequency regimes).
QPhox partnered with qubit company Rigetti to demonstrate optical frequency readout of superconducting qubits in 2023, an important first step towards enabling these qubits to be interfaced with a photonic network.
We are exploring potential collaborations with QPhox to investigate how transducers could be integrated with our networking technology. They have recently joined our Commercial Quantum Datacentre Alliance (confidential).

‍

Quantum memory companies

Lightsynq 🇺🇸

LightSynq is a spin-out of the Harvard academic team inside AWS Quantum Networks (Boston, USA), launched in November 2024. The company is led by Mihir Bhaskar as CEO, who was the PhD student at Harvard who led the experimental work on which the company was founded.
LightSync develops quantum memories based on atomic-scale impurities in nanosized diamonds. As part of their system they develop low-loss fibre-chip coupling and frequency conversion technology.

MemQ 🇺🇸

MemQ is a pre-seed startup founded by academics from the University of Chicago, established in 2021. They are building quantum memories using trapped ion qubits (erbium) on a silicon photonics platform, which operate at C-band telecom frequencies. Their use-cases include long-range quantum communications for a quantum internet, as well as distributed quantum computing and sensing applications.

Welinq 🇫🇷

Welinq is a spin-out of the University of Sorbonne in Paris (France), co-founded by Dr Eleni Diamanti, Prof Julien Laurat, and Dr Tom Darras. Established in 2022 and funded by the European Union and European Innovation Council, the company currently has a team of 16.
Welinq develops quantum memories based on neutral atoms, and has partnerships with other Paris-based start-ups Pasqal (neutral atom qubits) and Quandela (photonic qubits) to begin to integrate quantum memories with their partner’s processors by 2030. It is unclear if the required networking orchestration technology or qubit-photon interfacing technology will be developed by Welinq or their QPU partners.

‍

Quantum Memories - what can they be used for?

A quantum memory is any device which has at least 2 qubits, one to interface with light, and the other one to store a quantum state.

In our architecture, we use distributed local memory, native to each QPU. This means that whenever the QPI creates an entangled pair, that entanglement can be stored right there, using a qubit inside each of the entangled QPU, until it is needed for a computation.

There are a few startups proposing to use external memories to store entanglement between QPUs. These would be a third system that would sit inside each QNU, and would directly replace a passive optical component (the beamsplitter) inside the QNU for an active quantum memory.

What are the potential advantages of doing this?

In quantum networking, you interfere 2 photons coming from 2 different qubits at a beamsplitter (inside the QNU), as part of what is called a ‘Bell State Measurement’, which entangles the photons and hence the qubits that the photons came from.

Because photons need to arrive at the beamsplitter at exactly the same time, this introduces a (^2) term to the loss - so, the probability of entanglement success scales as 1/(Photon Loss)^2

If, instead of a beamsplitter, you have a ‘memory’, then you could remove the (^2) term, so:

Without memory: Entanglement success rate = (Qubit-Photon Entanglement Efficiency)/(System Loss)^2$,
With memory: Entanglement success rate =ubit-Photon Entanglement Efficiency)/(System Loss),

‘Qubit-Photon Entanglement Efficiency’ is the most crucial part of the equation, is the quality and rate of entanglement you start with, which is created at the QPI.

‘System Loss’ includes every single loss throughout the system: coupling the light into the fibre, loss through the fibre, through the switches, in the detectors, etc.

If the Qubit-Photon Entanglement Efficiency and the System Loss are not extremely good to begin with, then the ^2 factor doesn’t really matter much.

Nu Quantum increases by orders of magnitude the Qubit-Photon Entanglement Efficiency with our QPI technology, and we reduce significantly the System Loss with our quantum photonic chips. To our knowledge, the different startups producing quantum memories do not have the QPI or QNU technology to make the whole system. They would be a component supplier to a system integrator like Nu Quantum.

Could we integrate other companies’ quantum memories inside our QNU?

The answer is yes, with two major caveats:

Loss. In order for it to be worth it to integrate an extremely complex new system (a quantum memory) into our QNU, to replace en extremely simple system (a passive 50:50 beamsplitter), the gains you get because of the deleted ^2 factor need to significantly surpass the new losses that you introduce with the quantum memory. The loss in a beamsplitter is very low. So, the whole quantum memory system (which has tens of subsystems and components inside it) needs to be about the same. This is incredible challenging.‍
Fidelity. This is the real show-stopper. When we architect a datacentre-scale distributed quantum computing, we have very important targets on the fidelity of both local links (inside QPUs) and remote links (via the networks). We cannot tolerate anything which would affect the remote entanglement fidelity. Introducing a new coherent (quantum) component into our network, i.e. a memory, would almost certainly have a negative impact on the overall remote fidelity, which could render the whole system useless.

The other potential benefit of a quantum memory is that one wouldn’t need to synchronise the photons arriving at the beamsplitter so precisely, i.e. each one could arrive at a different time. We have done an analysis of whether photon synchronisation is something which would have a severe negative impact in our system and we don’t believe it is. One could always use a coiled up fibre to store photons for some time if needed, which would have the negative effect on fidelity and would be cheaper and easier to implement.

Conclusions on Quantum Memories

If in the future, quantum memory technology gets really good - i.e. 100% fidelity and almost 0 loss, them we could swap out the beamsplitters inside our QNU for their technology, and get an advantage on rate. Caveat: we would also need to do a cost analysis.
We believe quantum memory technology is important, but could perhaps find a better product-market fit as repeaters to enable long-distance quantum networking. This is a large and open market at the moment, where the performance targets on fidelity and rate are significantly relaxed compared to distributed quantum computing.
Nevertheless, we are keen to enable the quantum networking ecosystem and collaborate with credible quantum memory startups, and together could enable use-cases such as Blind Quantum Computation.

‍

No items found.