An IBM researcher holds a component of the Loon quantum computer.
IBM
Being a participant in the race to create error-free quantum supercomputerIBM is taking a different tack than its closest competitors. Now the firm has unveiled two new quantum computers, called Nighthawk and Loon, that could validate its approach and provide the innovation needed to make the next generation of these devices truly useful.
The design of the IBM quantum supercomputer is modular and is based on the development of new ways to connect superconducting qubits within and between different blocks of the quantum computer. When the company first introduced it, some researchers questioned the practicality of the connections, he says. Jay Gambet at IBM. He says it was as if people were telling the IBM team, “You're in theoretical land, you can't comprehend it.” AND 1762951424 we're going to show it [to be] wrong.”
In Loon, each qubit is connected to six others, and these connections can “split the plane,” meaning they don't just move around the chip, but can also move vertically—a capability that no other superconducting quantum computer has had before. Nighthawk, on the other hand, has four-way communication between qubits.
This increased connectivity could be the key to overcoming some of the biggest challenges facing existing quantum computers: it could increase their processing power and eliminate their error proneness. Gambetta says preliminary tests of Nighthawk show it can run quantum computing programs that are 30 percent more complex than those running on the firm's currently most used quantum computer. This increased complexity should lead to more applications for quantum computersand past IBM models have already begun to find application in areas for example chemistry.
However, the industry's Holy Grail is finding ways to group qubits into so-called logical qubits that are error-proof, and IBM is championing a method that requires those groups to be smaller than its competitors' approaches. for example the methods used by Google. This could allow IBM to achieve error-free calculations while avoiding some of the costs and engineering challenges associated with the need to create millions of qubits. But this doesn't work without a lot of connections between qubits, the type that Gambetta says they achieved with Loon.
Stephen Bartlett at the University of Sydney in Australia say that while more testing and benchmarking of new devices is needed, the expansion of qubit connectivity is exciting. “This is not a panacea that solves all the problems of scaling superconducting devices to the size needed for truly useful algorithms, but it is nonetheless a significant step towards that,” he says.
But there are still engineering and physical challenges ahead. One of these is developing a better way to read the output of a quantum computer at the end of a calculation, which Gambetta says is another area in which the firm has made progress recently. IBM Matthias Steffen says the team will now also work to increase every qubit “consistency time.” This is a measure of how long it remains in a quantum state useful for computation, which often degrades as new connections are added to the qubit. The team is also developing ways to reset some qubits during calculations.
In 2026, the firm plans to launch a modular quantum computer that can both store and process information, as reported by upcoming Loon and Nighthawk tests.
Topics:
