Article
Space Tech AI Models & Research

IBM links two cryogenic units in step toward fault-tolerant quantum computing

The company cooled and joined two modules, part of the plumbing needed for its 2029 Starling machine

by TechDefused Newsroom
The image showcases a close-up view of a quantum computer's intricate architecture. It highlights the complex arrangement of wires and metallic components, emphasizing the advanced technology used in quantum computing. — Credit: Photo by Planet Volumes / Unsplash cPhoto by Planet Volumes / Unsplash
Photo by Planet Volumes / Unsplash

IBM has connected and cooled two modular cryogenic units into a single shared environment, an early step toward much larger quantum computers.

Cryogenic systems keep quantum chips at temperatures close to absolute zero, the conditions they need to function.

The two modules cooled to below 15 millikelvin, more than 180 times colder than deep space, and together stand more than eight feet tall and wide.

The scaling problem

Scaling remains the central obstacle for superconducting quantum computers, the type IBM builds.

Individual processors can hold only so many qubits, the quantum equivalent of the bits in an ordinary computer.

Future machines will therefore need to link many chips inside one ultra-cold system.

Wednesday's result is a step toward that shared infrastructure, using a technology IBM calls the L-coupler to join separate processors.

The company plans to install its Nighthawk processors into the new modules later this year to expand testing.

The road to Starling

IBM is working toward Quantum Starling, which it says will be the world's first fault-tolerant quantum computer when it arrives in 2029.

A fault-tolerant machine can correct its own errors, the barrier that has kept quantum computers from useful work at scale.

Before then, IBM's roadmap targets a system with at least 1,000 programmable qubits by 2027, using the same L-coupler approach.

The announcement is incremental rather than a breakthrough, addressing the engineering of quantum computing rather than the qubits themselves.

But that unglamorous work is exactly what will determine whether IBM hits its 2029 goal.


by TechDefused Newsroom