IBM unveils updated quantum computing products

YORKTOWN HEIGHTS, NEW YORK — IBM is continuing its journey to scaling fault-tolerant quantum computers to commercial use with two new product launches that enable fast and accurate computation within their qubit architectures.

IBM announced on Wednesday that its new quantum processing unit, Nighthawk, and larger Loon lattice architecture are both slated to advance quantum computing’s benefits at two separate layers of the technology stack. 

“Our goal has always been to deliver fault-tolerant quantum computers, but to go on a path where we're always delivering our best platform,” Jay Gambetta, IBM Fellow and director of IBM Research, told reporters Nov. 7 at the company’s Yorktown Heights facility.

Nighthawk, a computing chip anticipated to be ready for use by the end of this year, features 120 qubits connected via 218 next-generation tunable couplers. This increased connectivity between qubits supports seamless data processing to solve intricate calculations that demand up to 5,000 two-qubit gates — the specific operations that manipulate quantum data.

IBM said it expects to continue expanding the number of gates Nighthawk can process to 7,500 by the end of 2026, with the goal of reaching 10,000 in 2027. 

In contrast, Loon is a larger, more complex processor, weaving circuits together to support improved chip connections for more reliable computing. Its novel architecture features longer couplers that physically connect qubits in a single lattice. As these couplers facilitate qubit connectivity, qubits have less barriers to entanglement, thus supporting stable quantum computing.  

“Loon is a device built to demonstrate a path to fault-tolerant quantum computing,” Gambetta said. “It requires many more layers of metal than Nighthawk, and it's our default path to build these future fault-tolerant quantum computers by 2029. So the key difference is that it has more couplers.”

The introduction of Nighthawk and Loon signals IBM’s push to reach a consistent level of quantum advantage through multiple avenues. The company’s quantum algorithm-specific library, Qiskit, is also getting an upgrade; a new C++ programming interface will expand its access to users working within high-performance classical computing environments. 

“We give the users all the freedom to tell the quantum computer how to operate,” Gambetta said. “So we're moving from just running quantum computers as circuits to running them as libraries.”

Qiskit’s expansion into high-performance programming languages like C++ mirrors fellow chip manufacturer NVIDIA in marrying classic high performance computing with emerging quantum computing capabilities to expedite and improve computation results. 

“At the end of the day, all that we care about when we have quantum advantage is: did my quantum plus classical thing run faster than my classical one alone? And now, with all these updates, we can trade quantum and classical in real time,” Gambetta said. 

IBM’s recent advancements have continued to garner federal attention as part of Stage B in the Defense Advanced Research Project Agency’s Quantum Benchmarking Initiative.