IBM has unveiled its latest and most powerful quantum computing chip to date, the 127-qubit Eagle processor.
Announcing the new piece of hardware at IBM Quantum Summit, it’s the first quantum chip developed by the company that surpasses 100 qubits.
Eagle is built using a new 3D packaging architecture developed by IBM which it says can support the development of future, more advanced quantum processors up to and including its proposed 1,126-qubit Condor chip, due for release in 2023.
The architecture is based on a heavy-hexagonal qubit layout – a setup that sees a qubit connecting to two or three neighbours. This configuration decreases the potential for errors caused by interacting neighbours and provides significant boosts in yielding functional processors, IBM said.
The architecture also places the qubits on a single layer while other components sit on different levels in a ‘stacked’ formation.
The processor will be available to select members of the IBM Quantum Network starting in December.
The IBM Quantum Network is a collection of Fortune 500 companies, academic institutions, startups, and national research labs that work with IBM to advance the field of quantum computing.
As part of IBM Quantum’s roadmap, Condor will mark a significant step in hardware advancements but further development depends on whether commercial dilution refrigerators can rise to the task of cooling such large, complex devices.
That said, the second announcement IBM made at its Quantum Summit may help towards that goal. The IBM Quantum System Two is designed to work with processors exceeding 1,000 qubits and will feature a more modular design with an overhauled cryogenic platform to optimise cooling performance.
IBM is on track to launch the system by 2023 which will help increase the scale of its chips.
The quantum computing capabilities of Eagle far exceed those of classical computers, Arvind Krishna, CEO at IBM, told HBO.
Classical computers encode information into bits represented as either a 1 or 0, while quantum computers encode information using a quantum superposition of a 1 and 0. This means information can be seen as representing a 1 or 0 – not both at the same time – a superposition is broken down to reveal a probability of revealing a 1 or 0.
The method of encoding allows quantum computers to process more complex tasks. Classical computers will work through problems in an order but quantum computers will approach problems differently, modelling all potential workloads and workstreams at once to generate answers much quicker.
It makes quantum computing ideal for working through large data sets and for tasks such as cracking cryptographic keys, for example.