French startup Pasqal launched in 2019 with an ambitious pitch: to build some of the world’s first quantum computers — devices based on a new computing paradigm, which scientists say could one day solve problems currently out of reach for traditional computers.
In doing so it put itself in direct competition with some of the most established and deep-pocketed tech companies in the world — in particular IBM, which had been working on developing the technology for years.
It’s now working hand in hand with the US tech giant as part of a partnership to advance ‘quantum-centric supercomputing’, a term coined in 2022 by Jay Gambetta, who has led IBM’s quantum efforts since 2018, to describe the integration of quantum and conventional computations within a single workflow to make the most of what each technology is best at.
Gambetta said IBM approached the French startup because it saw it as the most advanced competitor in the field.
“We’re both at a point where our hardware is beyond what you can brute-form simulate,” Gambetta told Sifted during the VivaTech conference in Paris last week. “There are many others building hardware who are not at that point.”
“Our hardware is now a unique piece of compute that you can use for things that cannot exist otherwise,” Pasqal’s CEO Loïc Henriet told Sifted.
In a sector as young as quantum computing, added Gambetta, the best way to move forward is for the top companies in the field to join forces.
“We’re building a new technology,” he said. “We’ve got to make an industry exist first. That’s step one.”
“Step two, obviously we’re going to compete on our own different hardware. But if you don’t start the race and get the industry in the first place, you can compete all you want — it’ll be for nothing.”
Developing algorithms
Cracking quantum-centric supercomputing requires building the software architecture and algorithms that enable the integration of quantum and classical workflows with different types of quantum computers.
Pasqal and IBM are effectively building different kinds of quantum devices. IBM’s computers are based on superconducting qubits, meaning the company uses electrons to create quantum bits — the tiny particles that carry quantum information inside quantum computers. Pasqal uses an approach called neutral atoms, in which specialised lasers trap single atoms to form qubits.
Pasqal and IBM researchers are working together to develop algorithms that could run across both types of hardware, as well as jointly with classical computers.
“I’ve said many times to everyone at Pasqal that I believe our hardware is the right hardware, and I’m sure Pasqal will say the same thing,” said Gambetta. “But hardware without algorithms is nothing.
“The biggest competition to everybody is that we don’t actually come up with more algorithms that allow us to reach the full potential of quantum […] So let’s make it easier for algorithm researchers to do their research.”
Quantum utility
Most quantum computing companies are currently focusing on growing the capabilities of their quantum processors, which means increasing the number of qubits they can handle.
Pasqal and IBM are closely aligned technologically, with both companies currently commercialising products with more than 100 qubits. Gambetta said this means they have passed the point of “quantum utility”, when a quantum computer can beat the simulations of quantum computing carried out by classical computers.
Unlike most other players in the sector, said Gambetta, Pasqal’s and IBM’s hardware are therefore “scientific tools” that can be used to build and test new quantum software.
“If you can simulate hardware on your laptop, you’re not doing any algorithm development,” said Gambetta.
IBM and Pasqal are not the only companies to have crossed the 100 qubit threshold. Although not a commercial product, a 105-qubit chip was announced by Google at the end of 2024, while Finnish startup IQM’s latest processors can support 150 qubits. Others are getting near: UK company Quantinuum, for instance, is set to introduce a system with 96 qubits.
The next step, which Gambetta expects will occur in the next two years, is to demonstrate quantum advantage — when a quantum computer beats a classical computer not only at simulating, but at solving an actual problem. “That’s a big deal,” he said.
In the medium term, however, both Pasqal and IBM are focusing on error-correction. This consists of systematically removing the errors that naturally occur when carrying out quantum computations, enabling the device to “become more like the ideal theory of a quantum computer,” said Gambetta.
Both IBM and Pasqal’s technological roadmaps point towards developing large-scale error-corrected devices by the end of the decade.
Read the orginal article: https://sifted.eu/articles/jay-gambetta-ibm-loic-henriet-pasqal-interview/
