IBM’s quantum roadmap aims to be one of the most ambitious. The company said last November, during its Quantum Summit, that the quantum advantage would be reached on certain applications as early as 2023, when the quantum processor will have reached one thousand qubits.
Jay Gambetta, vice-president of IBM Quantum, talks to ZDNet about the group’s strategic vision to bring quantum into the industrial dictionary.
Returning to this famous roadmap, IBM’s quantum boss said: “We try to focus on the things that we think are possible. In other words, “It’s still science, but I don’t see anything fundamental that says we’re not going to be successful.” I am convinced that we will implement each of these elements”.
In broad outline, remember that after deploying its first quantum computer in the cloud in 2016, IBM introduced in 2017 an open source software development kit for programming these quantum computers. Then, in 2019, the group arrived with its first quantum computer system, the IBM Quantum System One. At the same time, IBM delivered the Falcon processor (27 qubits) in 2019, then a year later the Hummingbird (65 qubits) and again a year later the Eagle processor (127 qubits). In 2022, “we are on the right track for Osprey (433 qubits),” Jay Gambetta told ZDNet, advocating a roadmap that is both “real and aggressive”.
A hybrid strategy for the industry
Aggressive, it is: IBM is indeed imagining 1,000-qubit machines for next year. Behind the roadmap, the group outlines a very concrete ambition. “We want to deal with real optimization problems in finance, for example, or even other use cases in AI and chemistry thanks to quantum technology”, underlines Jay Gambetta.
In sum, according to the vice president of IBM Quantum, “we want to ensure that quantum computing is useful to the industry and without friction. This means making classical and quantum work together in a hybrid strategy to create libraries to push the limits of what can be done.” But Jay Gambetta specifies that it is not a question of running on a quantum processor something that already works very well on a conventional processor. “It makes no sense,” he said.
And in this effort to find concrete cases of application, the industrial ecosystem has more than its word to say, believes Jay Gambetta. “I see a real industrial strategy that we are creating with many partners. Some industries will create certain parts of the software stack, others will take care of certain parts of the supply chain, and so on. »
IBM also nurtures premium relationships with some 180 partners in its Quantum Network. “It’s a way for industries and academics to be able to work together,” notes Jay Gambetta.
Access, R&D and workforce
For Jay Gambetta, it is also important to continue to invest in R&D and manpower to strengthen these ecosystems. To create vocations, the quantum manager insists in particular on university training and the potential for retraining of expert profiles in mathematics.
To sum up, the industrial strategy is therefore based on four pillars, which are “access, R&D, labor and industry”, indicates Jay Gambetta.
On access, the manager also notes that other Quantum System One machines should be deployed at the Cleveland Clinic in the United States, Quebec and South Korea. Quantum systems are already deployed outside the United States, including one in Germany and another in Japan.
For France, Jay Gambetta observes a fairly dynamic quantum ecosystem “around promising innovations and interested industries”. While he believes there is room for several technologies other than IBM’s, such as Pasqal’s neutral atoms or Quandela’s photonics, he encourages start-ups to “invest in hardware” to be to the state of the art without widening the gap.