According to a report by consultancy McKinsey published in April, quantum computing has ‘reached a commercial tipping point’. Companies in sectors such as climate, life sciences, transport and logistics, or financial services are already using this emerging technology to optimize processes, test scenarios and run simulations.
Investors also poured $12.6 billion into quantum startups over the previous year, another sign of commercial interest in this technology. For John Roese, global chief technology officer and head of artificial intelligence at Dell Technologies, the question is no longer whether the thing is possible, but when it will become commonplace. It will not be a sudden tipping point, but a steady progression towards fault-tolerant systems, i.e., capable of correcting their errors, and increasingly usable.
Infrastructures being put in place
Major server manufacturers are already preparing the ground. Dell, Hewlett Packard Enterprise, Cisco and IBM are installing the infrastructure components needed for hybrid data centers, combining quantum computing and classical high-performance computing. Cisco is developing a quantum-dedicated network architecture, while Dell, HPE and others are working with quantum technology providers. At Dell Technologies World 2026, the company Equal1 presented its RacQ system.
HPE for its part announced partnerships with several quantum players, including IQM Quantum Computers, Quantinuum, QuEra and Rigetti, to integrate their systems into the company’s Cray supercomputers.
Large-scale optimizations
According to John Roese, users will see the first effects in the short term. ‘Before the end of this decade, so around 2028 or 2029, we will probably have enough error-corrected logical qubits to tackle the first batch of real-world problems, which are essentially optimizations at the molecular level,’ he said. A qubit, or quantum bit, is the basic unit of quantum information, and a logical qubit groups several physical qubits to make the calculation reliable.
At the beginning of the next decade, systems should enable a second step, dedicated to larger-scale optimizations, such as those that airlines or financial companies must coordinate. ‘The entire global supply chain is fundamentally sub-optimal, because we don’t have the computing power to perform large-scale optimization calculations,’ he said. He also mentions the number of qubits needed: a 1,000-qubit system would not be enough to model all the roads in the United States, and it could take ‘a million, or even several million’.
Architectures developed by Qarakal Quantum and Infleqtion aim to reduce the number of physical qubits needed to create a logical qubit, which would help limit infrastructure costs and energy consumption.
In the molecular domain, Mr. Roese cites the example of corrosion. He points out that the orientation of atoms determines the properties of materials, as in stainless steel, and that better control of these structures could enable the design of more resistant materials. ‘The numbers I’ve seen indicate that corrosion costs the global economy a trillion dollars a year,’ he added. Batteries and drug discovery also fall into this type of optimization.
For logistics, D-Wave offers quantum annealing systems, specialized in optimization, and is also working on superconducting systems. Mr. Roese describes the ‘airport gate problem’: with 30 gates, 300 flights and 100,000 passengers, finding the best combination becomes very difficult to calculate. He also cites FedEx, UPS and DHL, which use route optimization formulas to limit their drivers’ left turns.
Finally, Mr. Roese anticipates that by 2030, the number of autonomous artificial intelligence agents will increase sharply. ‘A trillion or something like that?’ he wonders, stressing that once assembled into systems, these agents become difficult to understand, control and direct.
