Optimization Is the Key Challenge for Quantum Computing

Quantum Computing Moves Toward Practical Optimization

A McKinsey & Co. report published in April said quantum computing had reached a commercial tipping point. Companies in climate and life sciences, travel and logistics, and financial services are already using the technology to optimize processes, test scenarios, and run simulations. Investors put $12.6 billion into quantum startups last year, providing another indication of growing commercial interest.

These developments suggest that quantum computing has moved beyond asking whether it can be built. The focus is increasingly on when practical, scalable systems will become available. John Roese, global chief technology officer and chief AI officer at Dell Technologies, describes this as a gradual progression toward fault-tolerant systems with improving error-corrected qubits, architectures, software, and applications.

Companies including Dell, Hewlett Packard Enterprise, Cisco Systems, and IBM are developing infrastructure for hybrid quantum-classical HPC datacenters. Cisco is working on a quantum-focused networking architecture, while Dell, HPE, and other companies are partnering with quantum vendors. At Dell Technologies World 2026, the quantum company Equal1 displayed its RacQ system.

HPE has also announced partnerships with quantum companies including IQM Quantum Computers, Quantinuum, QuEra, and Rigetti. The goal is to integrate their systems with HPE Cray supercomputers.

Early Applications Will Focus on Molecular Problems

McKinsey’s data indicates that early adopters are already seeing benefits. Those benefits, and the number of organizations gaining from them, are expected to increase as quantum systems become more capable.

Roese said users could begin seeing meaningful results before the end of the decade. By around 2028 or 2029, he expects quantum systems to have enough error-corrected logical qubits to address the first group of practical problems, particularly molecular-level optimization.

By the beginning of the next decade, quantum systems could have sufficient capacity for larger optimization problems. Potential applications include coordinating airline and transportation operations and managing complex financial decisions. These problems involve many interacting variables and can be difficult for even powerful classical computers to process within practical time and cost limits.

The main constraint remains the number of available qubits. Roese noted that representing a large data set, such as the roads across the United States, would require far more capacity than a 1,000-qubit system provides. A large optimization problem could require one million or many millions of qubits.

Architectures from companies such as Qarakal Quantum and Infleqtion are intended to reduce the number of physical qubits needed to create each logical qubit. Reducing that ratio could lower infrastructure and energy requirements, reduce noise, and make it easier to place quantum systems alongside classical HPC equipment in datacenters.

Even early systems could produce results with practical effects, including at the molecular level. The performance of batteries and other materials often depends on how atoms are arranged within their molecules. Optimization involves more than simply connecting atoms. Their spacing, orientation, and binding characteristics also affect the resulting material.

Roese used corrosion as an example. At a basic level, rust can result from molecular structures that are not optimally arranged. Stainless steel resists corrosion because of the arrangement of its atoms and the properties that arrangement produces. Quantum systems could eventually help researchers identify why materials corrode and adjust their structures to make them more resistant. Roese cited estimates that corrosion costs the global economy about $1 trillion annually.

Traditional supercomputers can address some of these questions, but quantum systems are expected to handle certain calculations faster and at a larger scale. Drug discovery and the development of longer-lasting batteries are other examples of molecular-level optimization. Quantum computing is therefore becoming an engineering challenge as well as a physics challenge, with applications whose effects may be largely invisible to consumers while still influencing many industries.

Logistics Creates Larger Optimization Problems

By 2030, quantum systems could begin tackling optimization problems beyond the molecular scale. Roese identified logistics as one of the largest opportunities.

D-Wave’s quantum annealing systems are designed specifically for optimization problems, although the company is also developing superconducting gate-model quantum systems. Dell has used annealers for supply-chain optimization. However, annealers are specialized for particular optimization workloads, while broader quantum systems could address more varied problems.

Airline gate scheduling illustrates the challenge. An airline might need to assign 300 flights to 30 gates while coordinating 100,000 passengers and accounting for different aircraft configurations. The number of possible combinations grows rapidly, making the problem difficult to solve manually or with conventional methods within a reasonable time.

Transportation companies such as FedEx, UPS, and DHL provide another example. Their route-optimization systems can reduce the number of left turns drivers make, saving time and fuel. This is a relatively simple form of optimization. A more comprehensive model would examine all possible routes, traffic conditions, delivery requirements, supply sources, and other dependencies to identify a better overall plan.

Quantum optimization could be applied across multiple layers of such systems. Delivery vehicles represent one layer, while the goods they transport may come from different sources and move through several stages before reaching customers.

Preparing for More Autonomous Systems

Optimization challenges will become more complex as AI agents and autonomous vehicles become more common. By 2030, there could be vastly more autonomous agents operating in both physical and digital environments.

Roese said individual autonomous systems may be predictable in isolation but become difficult to understand and control when they operate as teams. A large ecosystem of autonomous AI entities could contain too many interacting variables for organizations to model or simulate effectively with existing systems.

A sufficiently capable quantum system could eventually help coordinate these entities by modeling their interactions and identifying effective courses of action. The workloads addressed after the initial molecular and larger-scale optimization phases are expected to increase in both number and complexity.