Quantum computing - a long race that is far from decided

Quantum computing is still in its infancy. Most vendors are still in pure research mode, but some have introduced the first usable quantum computers for on-premises use or, more commonly, as a cloud service.

There are two fundamentally different approaches:

• Quantum annealing: useful for optimization issues such as route optimization (traveling salesperson problem), but not for general use cases.
• Gate-based quantum computers: with the potential for quantum supremacy (Quantum supremacy refers to the superiority of quantum computers over classic supercomputers in solving complex problems. This refers to the point in time when a quantum computer can solve a task in an acceptable period of time for which a computer based on conventional digital technology would require unfeasible computing time) when it comes to general use cases, mainly for AI and machine learning, financial modeling, cybersecurity, route and traffic optimization, realistic prototyping and testing in manufacturing, as well as drug, material, and chemical research.

As current research is looking for general-purpose quantum computers, the future of quantum computers will be gate-based systems. For those systems, several qubit implementations are used in gate-based quantum computers, each with its own performance tradeoffs:

• Superconducting qubits: recognized technology option with significant investment and research which, however, requires extremely low temperatures to function. Relevant vendor examples are IBM and Oxford Quantum Circuits.
• Trapped ions: long coherence times, resulting in higher qubit stability but also scalability issues due to the physical limitations of ion trapping. Relevant vendor examples are Quantinuum, IonQ, and Alpine Quantum Technologies.
• Photonic qubits: operates at room temperature, so no ultra-low temperature environment is required; however, photonic components can be relatively large, making miniaturization difficult. Relevant vendor examples are Xanadu Quantum Technologies and Quandela.
• Neutral atoms: long coherence times and high gate fidelity, but also slower gate operation times than superconducting qubits. Relevant vendor examples are PASQAL, Atom Computing, and QuEra Computing.
• Semiconductor/silicon points: inherently fault-tolerant due to the nature of topological quantum states; however, the technology is still at an early stage of development. Relevant vendor examples are Intel and Google.

It remains to be seen which basic technology will ultimately prevail in quantum computing. Superconducting qubits and semiconductor/silicon points stand a good chance, but the other approaches cannot be ruled out either. That is why user organizations and IT service providers should currently be technology agnostic. Until the race is decided, using quantum computing in a cloud computing model is the best option. Large companies might one day want to run a quantum computer on-premises, but not at this stage.

Recommendations for providers

• The first commercial offerings show a clear business value for user companies with actional needs. Prepare to support these customers.
• Make sure to build internal knowledge about quantum computing and, in particular, use cases, algorithms, and code development.
• Collaboration with research institutions and leading manufacturers of quantum computers can be beneficial.
• Many users will use quantum computing as a cloud service but still need support on the application side. This should be a focus from the start.