Quantum computing firm Quantinuum has brought forward the date when it believes quantum computers will become commercially viable. A study from McKinsey estimates that 5,000 quantum computers will be operational by 2030 but it may be 2035 before hardware and software is readily available to solve computationally complex problems.
Quantinuum’s technical roadmap shows that by 2029, the company will be producing a machine called Apollo that has thousands of physical qubits, which translates to hardware with hundreds of logical qubits and claimed “physical error rates less than 10-4”, according to Quantinuum.
In a blog post, Quantinuum’s chief scientist for algorithms and innovation, Harry Buhrman, and Quantinuum fellow Chris Langer discussed how Apollo could use up to 10,000 gates for less complex computational tasks and apply hundreds of qubits for longer computations.
“By the end of the decade, our accelerated hardware roadmap will deliver a fully fault-tolerant and universal quantum computer capable of executing millions of operations on hundreds of logical qubits,” they wrote.
The first application areas for Apollo are expected to be in scientific discovery for the simulation of quantum systems. According to Buhrman and Langer, application areas include discoveries in materials science, high-temperature superconductivity, complex magnetic systems, phase transitions, and high-energy physics, among other things.
The pair said they have brought forward the roadmap thanks to three technological breakthroughs. The first is what they call “the wiring problem”, whereby control signals need to be sent to each qubit to perform the necessary operations required for a computation, which means the number of control signals increases in line with the number of qubits. This, according to Quantinuum, is both impractical and prohibitively expensive. To get around this, Quantinuum has developed a protocol that broadcasts shared control signals with qubits arranged in a 2D geometry.
The second development is the company’s work to reduce what Buhrman and Langer call “two-qubit physical gate errors”.
Third is what they describe as “all-to-all” connectivity. This was recently demoed with Microsoft, where 56 physical qubits were used to generate 12 logical qubits. Buhrman and Langer said the work demonstrated several experiments, including repeated rounds of error correction where the error in the final result was approximately 10 times lower than the physical circuit baseline.
Before the company can get to Apollo, it will be releasing Quantinuum Helios, powered by Honeywell, a device it says will be able to move beyond classical computing capabilities in 2025.
In 2027, the company plans to introduce Quantinuum Sol, its first commercially available 2D-grid-based quantum computer. Quantinuum said Sol will offer hundreds of physical qubits and operate approximately twice as fast as Helios, with a much higher level of error correction. “Sol, being a fully 2D-grid architecture, is the scalability launching point for the significant size increase planned for Apollo,” wrote Buhrman and Langer.
“Apollo promises fault-tolerant quantum advantage sooner, with fewer resources,” they added.