Today, IBM unveiled two new quantum computers. The bigger of the two, dubbed Condor, is the second ever to have a total number of quantum bits, or qubits, in the quadruple digits. IBM鈥檚 other new quantum computer is called Heron, and it is the company鈥檚 least error-prone device to date.
Researchers disagree on how to best build a quantum computer, but there is broad consensus that quantum computers will eventually be able to solve problems that are impossible even for the most powerful conventional computers. While Atom Computing, the company behind the current largest quantum computer, and IBM have now developed similarly-sized devices, their qubits are built differently. Atom Computing鈥檚 relies on qubits made from neutral atoms, while IBM bases its qubits on tiny circuits that conduct electricity without any resistance.
In 2022, IBM broke the record for the largest quantum computer with a 433-qubit device. Engineers made it work by improving the way its qubits were controlled. at IBM says the team had to overcome similar technical challenges to build Condor.
Advertisement
鈥淐ondor solves a scaling problem. It shows that we can get to 1000 qubits, and that sets us up for the future,鈥� he says. Condor has 1121 qubits, just 59 fewer than Atom Computing.
To make Condor, the IBM team focused on improving the quantum computer鈥檚 input mechanisms and how its output is read. All of these processes required connecting the quantum computer to conventional electronics. This is challenging, because if those connections are not carefully engineered, they can destroy the special quantum properties that enable such devices to work at all.
鈥淏reaking the barrier of 1000 qubits by any company is a major milestone for the quantum computing industry,鈥� says at Atom Computing. However, this number alone does not tell the whole story. Factors like the stability and reliability of qubits, as well as how fast they work, also matter, he says.
Gambetta says building the much smaller Heron, which has 133 qubits, helped IBM confront some of those challenges. Quantum computers are prone to making errors, some of which stem from unintended interactions between qubits, known as 鈥渃rosstalk鈥�. The researchers designed Heron鈥檚 hardware to make it as immune to this effect as possible by creating a more controlled way for qubits to exchange information.
According to IBM, the rate at which Heron makes errors is five times smaller than previous state-of-the-art quantum computers. This sets it up to run longer and more complex programs and allows researchers to experiment with finding the best quantum computing programs for different problems, says Gambetta.
鈥淭he most disruptive applications for quantum computers are sitting firmly within the million or many millions of qubits domain,鈥� says at the University of Sussex in the UK. He says at that scale, quantum computers will become fault tolerant 鈥� they will be able to catch and correct their own errors, making them useful in industries ranging from pharmacy to aviation.
鈥淲hat can a 1000-plus qubit quantum computer be used for in the meantime? At this stage quantum computers really are research tools,鈥� says Bloom. Academic and industry developers will use these platforms to explore increasingly complex problems that scale with the number of qubits.
But IBM is betting on improvements like tweaked algorithms and better error mitigation to push its existing fleet of devices towards more immediate utility. In June, the company鈥檚 researchers showed that a quantum computer can beat a conventional supercomputer in a race to complete some complex calculations. However, this feat was followed by a slew of calculations proving there are still ways for the supercomputer to win.
