MilestoneQuantum advantageDisputed

Google's D-Wave speedup claim sparks debate over fair comparisons

Dec 8, 2015

Google reports a 100-million-fold speedup on specially designed problems using a D-Wave 2X, compared with simulated annealing on one CPU core. Other classical algorithms erase the advantage, so the result does not establish a general quantum speedup.

Why it matters

A quantum machine must beat the strongest relevant classical methods, not just one selected benchmark, to demonstrate a useful advantage.

The benchmark result is not a general speedup over classical computing. Scott Aaronson notes that both reported advantages disappear against other classical algorithms; Google's own announcement acknowledges faster classical solvers for these tests.

Google · D-Wave · NASA

See it on the timeline

Google and NASA launch a quantum AI lab built around a D-Wave machine (May 16, 2013)
Google claims quantum supremacy with its 53-qubit Sycamore processor (Oct 23, 2019)

Sources (3)

Google Research When can Quantum Annealing win?
Primary sourceDec 8, 2015

“We recently applied these new insights to construct proof-of-principle optimization problems and programmed these into the D-Wave 2X quantum annealer”
arXiv (Denchev et al.) What is the Computational Value of Finite Range Tunneling?
Primary sourceDec 7, 2015

“$\sim 10^8$ times faster than SA running on a single processor core.”
Scott Aaronson (Shtetl-Optimized) December 2015 discussion of Google's D-Wave results
Challenges the claimDec 2015

“And in any case, both speedups disappear if you compare against other classical algorithms, like that of Alex Selby.”

Added Oct 19, 2026. Spotted an error? Tell us and we will review it.

Google's D-Wave speedup claim sparks debate over fair comparisons

Why it matters

Related