Let’s start at the beginning of the field. In 1981, Feynman shared this idea: To simulate quantum phenomena, use machines based on quantum mechanics1. Instead of bits, use qubits. Over the next ~15 years, physicists and computer scientists defined these machines. In particular, Deutsch, Josza, Simon and Vazirani showed how machines based on quantum mechanics could solve some problems faster than classical computers.
Arguably the most influential event in quantum computing was in 1994, when Shor showed that these machines were useful for something other than contrived problems and simulating quantum mechanical systems. Shor’s algorithm proved that quantum computers could eventually break RSA and most of public-key cryptography. This moved quantum computers from the realm of interesting-to-physicists to potentially-important-for-lots-of-people.
A NIST conference was organized to discuss Shor’s Algorithm, which led to DARPA and the NSA funding research into quantum computers and algorithms. It wasn’t clear how long it would take to build them. Trapped ion qubits were proposed in 1995 by Cirac and Zoller, and realized almost immediately. There were basic NMR qubits too, but it seemed clear that they wouldn’t scale well. By the mid 2000s, IBM had ‘factored’ 15, DARPA was running a quantum key distribution network and Microsoft had launched Station Q. Governments, funding agencies and research-minded corporations were interested in QC.
Then, in 2012: ‘quantum supremacy’. John Preskill coined the term to mean something very reasonable. IBM, D-Wave, Google and their comms teams contorted it until we all agreed never to use it again. Labs started forecasting the date by which they would achieve quantum supremacy, similarly to AI CEOs competing on how early their models will win a Nobel Prize, or automate software engineering.
Google acquired a world-class quantum computing team in 2014, by buying John Martinis’ group at UC Santa Barbara. That same year, the Snowden files revealed that the NSA had a quantum computing project. Startups materialized around this time too: Rigetti in 2013, IonQ in 2015, PsiQuantum in 2016. IBM had been working on quantum computers already, as had the national labs, but there was suddenly a commercial race, and the press releases ratcheted up in intensity.
I suspect that Google entering the race in 2014 was to quantum computing as spinning out Waymo in 2016 was to self-driving cars. At least one reason for the excitement was that qubit count appeared to have hit an inflection point, going from 9 qubits in 2016, to 17 qubits in 2017 and 72 qubits in 2018. In an effort to outdo Moore’s Law, “Neven’s Law” appeared for quantum computing. Most companies released aggressive roadmaps with targets for 2020 that they still haven’t hit, in 2025.
In 2018 and 2019, Google and IBM demonstrated quantum supremacy. It was a huge moment, something like an LLM solving a math theorem, and similarly controversial. Uninhibited press releases and media coverage implied that classical computing was finished. Online debates raged over whether or not the experiments actually constituted quantum supremacy. Despite all this, quantum supremacy did not bring quantum computers any closer to being useful, though it might have helped with funding further research. The debate was more of a PR battle between Google and IBM, with Google winning. In contrast to the LLM analogy, the underlying technology (quantum computers) was not improving fast enough for this technical milestone (quantum supremacy) to signify any impact on normal people.
In tech circles, things grew quiet. Progress in qubit counts slowed down. A lot of skeptics probably felt vindicated, though governments and corporations continued talking about ‘quantum’ and including it in their roadmaps, innovation plans, strategic alignment committee guidelines etc. And the areas they funded! Some of the QC companies went public, with IONQ and Rigetti going doing so via SPAC. Take from that what you will.
Researchers continued making steady progress, even if people were not watching as closely. Qubit fidelities hit 99%, and teams working on neutral atoms caught up with superconductors and trapped ions. With the announcement of Google Willow in December 2024, a new PR cycle kicked off, with headlines just as aggressive and overstated as in the past. This time around, retail investors are participating via the publicly traded quantum companies and some companies are playing into it, claiming “QPUs will ultimately start to replace GPUs.” I don’t know when aggressively optimistic salesmanship becomes fraud, but some people are doing well: Rigetti is up 30x, IonQ is up 4x. This will hardly last.
Hype about quantum computing is nothing new. Few people have any incentive to do something about it: Founders and investors want their companies to do well, management teams can sell stock for enormous payouts, governments and corporations don’t want the embarrassment of backtracking on loudly announced ‘quantum roadmaps’, and academics don’t want to endanger funding for their research. Short sellers are starting to criticize the statements, as are many skeptics, but I think they are much weaker forces. Perception of QC companies can get totally detached from reality. It might correct after some companies miss whatever roadmap targets they have recently announced - or not!
Despite all this, real progress is being made in quantum computing. I’m optimistic that 100,000-qubit machines might come online in 5 - 10 years’ time, most likely via neutral atom architectures.