Quantinuum, a quantum computing company, has revised its timeline for when quantum computers will achieve commercial viability. According to a study by McKinsey, it’s anticipated that by 2030, 5,000 quantum computers will be in operation, but practical hardware and software capable of addressing complex computational problems may not become available until 2035.
The company’s technical roadmap indicates that by 2029, it will launch a machine named Apollo, equipped with thousands of physical qubits. This is expected to result in hardware with hundreds of logical qubits and “physical error rates less than 10^-4,” as noted by Quantinuum. In a blog post, Chief Scientist for Algorithms and Innovation Harry Buhrman and Fellow Chris Langer elaborated on how Apollo could utilize up to 10,000 gates for simpler computational tasks and leverage hundreds of qubits for more extensive calculations. “By the end of the decade, our accelerated hardware roadmap will produce a fully fault-tolerant, universal quantum computer capable of performing millions of operations on hundreds of logical qubits,” they stated.
Apollo’s initial application areas are predicted to be in scientific discovery, particularly for simulating quantum systems. According to Buhrman and Langer, potential applications include advancements in materials science, high-temperature superconductivity, complex magnetic systems, phase transitions, and high-energy physics, among other fields.
Buhrman and Langer attributed the expedited roadmap to three key technological advancements. The first is the resolution of what they term “the wiring problem,” which involves sending control signals to each qubit for computation. This issue leads to a significant increase in control signals as the number of qubits rises, making it impractical and costly. To address this, Quantinuum has developed a protocol that broadcasts shared control signals to qubits arranged in a 2D layout.
The second breakthrough involves the company’s efforts to minimize “two-qubit physical gate errors,” while the third pertains to achieving “all-to-all” connectivity. This was recently demonstrated in collaboration with Microsoft, where 56 physical qubits were utilized to generate 12 logical qubits, showcasing several experiments that included multiple rounds of error correction, resulting in a final error rate approximately ten times lower than the physical circuit baseline.
Before the launch of Apollo, Quantinuum plans to introduce Quantinuum Helios, powered by Honeywell, in 2025, a device expected to surpass classical computing capabilities. In 2027, the company aims to roll out Quantinuum Sol, its first commercially available 2D-grid-based quantum computer, which will feature hundreds of physical qubits and operate about twice as fast as Helios, with enhanced error correction. “Sol, with its fully 2D-grid architecture, serves as the foundation for the significant scalability planned for Apollo,” Buhrman and Langer noted. “Apollo anticipates delivering fault-tolerant quantum advantages sooner, utilizing fewer resources,” they added.
