IBM has introduced the first quantum computer with more than 1,000 qubits, the equivalent of digital bits in a regular computer. But the company says it will now shift gears and focus on making its machines more resilient to errors rather than making them larger.

For years, IBM has been pursuing a quantum computing roadmap that roughly doubles the number of qubits each year. The new chip, called Condor and introduced on December 4, has 1,121 superconducting qubits arranged in a honeycomb pattern. This follows other record-breaking quantum processors, including a 127-qubit chip in 2021 and a 433-qubit chip last year.

For those who don’t know, quantum computers promise to perform certain calculations that are inaccessible to classical computers. They do this by exploiting unique quantum phenomena such as entanglement and superposition, which allow multiple qubits to exist in multiple collective states at the same time.

Threshold exceeded in quantum processors

However, these quantum states are notoriously unstable and error-prone. Physicists are trying to get around this by convincing several physical qubits, each encoded in a superconducting circuit or a single ion, to work together to represent an information qubit or ‘logical qubit’. As part of its new approach, the company also introduced a chip called Heron, which has 133 qubits but has a record-low error rate, three times lower than its previous quantum processor.

Researchers generally say that the most advanced error correction techniques will require more than 1,000 physical qubits for each logical qubit. In this case, a machine that can perform useful calculations must have millions of physical qubits. But in recent months, physicists have introduced an alternative error correction scheme called quantum low-density parity check (qLDPC). This scheme promises to reduce that number by a factor of 10 or more, according to a preprint prepared by IBM researchers. The company says it will now focus solely on building chips designed to hold a few qLDPC-corrected qubits within 400 or so physical qubits and then connecting those chips together.

some difficulties

Mikhail Lukin, a physicist at Harvard University, says the IBM preprint is an “excellent theoretical work” but claims that implementing the idea is difficult and even the first proof of concept can take years. Lukin and his colleagues had conducted a similar study on the possibility of implementing qLDPC using individual atoms instead of superconducting loops.

The problem is that the qLDPC technique requires each qubit to be directly connected to at least six others. In typical superconducting chips, each qubit is connected to only two or three neighbors. But Oliver Dial, IBM’s condensed matter physicist and chief technology officer of IBM Quantum, says the company has a plan: A layer will be added to the design of quantum chips to allow for the extra connections required by the qLDPC scheme.

The new roadmap for quantum research that IBM announced today predicts achieving useful calculations, such as simulating the functioning of catalyst molecules, by the end of the decade. Dial expresses that this is tremendous with the following words: “This was always a dream and always a distant dream. In fact, it is a tremendous thing for me to be so close that we can see the road from where we are today.”

We’re getting closer to quantum reality

IBM also announced the launch of Quantum System Two, a modular quantum computer. The system, located in New York, is already up and running and is initially powered by three Heron processors. According to IBM, we are currently in the 2nd era of quantum computing and efforts need to be made to reduce and mitigate errors and develop proof-of-concept applications. The company plans to build eight quantum computing centers where researchers can access System Two.

However, in the 3rd era, quantum computers will provide error correction and be scalable. IBM also aims to make the development of quantum computing accessible to a wide range of users and is developing Qiskit, a software stack that will allow developers to create code for a variety of applications. IBM’s press release stated that Qiskit Patterns will enable users to develop, deploy and execute workflows in both classical and quantum computing environments.

IBM’s roadmap includes processors called Flamingo, Crossbill and Kookaburra, which, once developed, could be combined today to create Condor’s 1000+ qubit capacity. However, IBM has set some milestones for the development of quantum computing systems, and given its track record, it looks like the company will make quantum computing a reality within the next decade.