This post is an attempt to deep dive into the computing systems built by D-Wave. D-Wave was the first company to build a commercially viable quantum annealer, in 2011. Although the company has spent most of its time working on annealing computers, they did announce in 2021 that they were working on a universal gate computer as well.¹

The company was founded in Burnaby, British Columbia, Canada in 1999, by Haig Farris, Geordie Rose, Bob Wiens, and Alexandre Zagoskin. The company name comes from a type of superconductor, known as a d-wave. The first actual public demonstration of their quantum annealing computer was at the Computer History Museum, February 15th 2007. The prototype system was limited to 16 qubits and had the codename “Orion”.²

The first commercial machine that was launched in 2011, known as D-Wave One, had a 128-qubit processor, which can perform a single mathematical operation, discrete optimization. The QPU codename was Rainier, as seen in the image below.

Soon after the announcement, it was made public that Lockheed Martin signed a contract to purchase one of these computers. Delivery for the new computer happened in October 2011, at USC’s Information Sciences Institute (ISI).

The next big step for the company was the launch of the D-Wave Two system, with a 512 qubit processor, code name Vesuvius in 2013. Later in the year a collaboration between NASA, Google and the USRA purchased one of the new system to be installed at the Ames Research Center in California. Around the same time there was a paper published by a D-Wave consultant comparing the speed of the D-Wave system to that of a classical computer. This study looks at three different pieces of software on a conventional computer vs. the D-Wave system; CPLEX, TABU running on METSlib and AK. The paper presented 3 instances of NP-Hard optimization problems;

“In all three experiments the V5 hardware or its hybrid counterpart Blackbox found solutions that tied or bettered the best solutions found by software solvers. Performance was especially impressive on instances that can be solved directly in hardware. On the largest problem sizes tested, the V5 chip found optimal solutions in less than half a second, while the best software solver, CPLEX, needed 30 minutes to find all optimal solutions.”³

A few years later, in 2015 they launched the D-Wave 2X machine, with ~1000 qubits. Due to manufacturing variability each QPU would have between 1152 and 1000 qubits available for use. At this point in time D-Wave switched from trying to say their machine benefited from quantum speed up, and focused on the performance of their QPU compared to other classical single threaded CPUs. Sometime in 2015 the Los Almos National Laboratory acquired a D-Wave 2X machine as well. When the next QPU was available, the D-wave 2000Q, both NASA and Los Almos received an upgrade to the new 2000 qubit chips sometime later around 2019.

In February 2019 D-wave announced the next installment in their QPUs for the quantum annealers, the Advantage, when it finally released as a commercial product in October 2021, it had over 5,000 qubits.

June 2022 the company presented the Advantage2 system as a prototype with 500+ qubits, with a roadmap up to an expected 7,000 cubits. Technical deep dive report here here.

September 2022 there was a joint paper released from D-Wave and Saitama Medical University (Japan) on an experiment run on the Advantage computer with up to 2,000 qubits. They were able to demonstrate the ability compute coherent quantum dynamics in large scale optimization problems.

April 2023, a medium article from D-Wave talks about the Advantage system with up to 5000 qubits. This also came with the paper in Nature about the 5000 qubit computer experiment which uses the Advantage QPU.

Most recently, In January 2024 the company announced its new Advantage2 prototype chip that provides 1,200+ quibits. This new chip is said to be 20 times faster than the previous generation Advantage system. When the Advantage2 system is fully launched it will be the companies 6th generation quantum annealer.

There has been a lot of controversy surrounding D-Wave since their initial launch, and it seems like no one has been able to conclusively prove one way or another if their computers are truly any faster than modern computers. Back in 2014 there were published papers showing that the D-Wave machines were not faster than a normal computer.⁴ Although the papers published in 2022 and 2023 by D-Wave seem to state they have see an advantage in some cases.