Can we have an honest conversation about where quantum computing actually is in 2026. Not the press release version.

[DQ Eric]

I have been following this field for four years and I am increasingly frustrated by the gap between what gets announced and what is actually happening. Not because progress is slow, it is not, genuine things are happening. But because the press release version and the engineering reality are so different that it is hard to have a sensible conversation about either.

The honest 2026 position as I understand it is this. We are in the NISQ era, meaning devices exist with 50 to around 1000 qubits but error rates are still high enough to limit circuit depth significantly. Google's Willow chip did something genuinely impressive in December 2024, completing a benchmark task that would take the best classical supercomputer longer than the age of the universe. But that benchmark was designed to be hard for classical computers, not to solve a commercially useful problem. The distinction matters enormously and most coverage collapses it. The QuEra 2026 Quantum Readiness Report found that 46 percent of organisations expect flat quantum budgets this year. The show-me phase is real.

The bit I find most interesting and least discussed is what a Caltech researcher said about algorithm readiness. If given a perfect quantum computer tomorrow, most of the field could not actually implement their quantum algorithms on it. The hardware is racing ahead of the theoretical and software ecosystem. There are estimated to be only 1800 to 2200 people worldwide specialising in quantum error correction. That is the entire global talent pool for the most critical technical area. McKinsey found there is one qualified quantum candidate for every three open roles.

Fault tolerant quantum computing, the era where you can run deep algorithms reliably, is most honestly a 2029 to 2032 story under optimistic but evidence grounded assumptions. Broad commercial transformation of industries is a 2040s story. Those timelines are not pessimism, they are what the physics and engineering are actually pointing at. The field has broken through the error correction wall, which one researcher called escape velocity, but the distance to destination is still significant

[VoidSentinel74]

Thank you for writing this. The gap between what IBM announces at a keynote and what the Qiskit documentation tells you is possible is genuinely huge and nobody seems to talk about it

[SpinState]

The algorithm gap is the thing that gets skipped every time. You can have a hundred logical qubits and still not have anything useful to run on them because the algorithmic development has been neglected

[Holly]

That is a bit strong. Quantum chemistry algorithms are reasonably well developed. The problem is more that the hardware is not yet good enough to run them at meaningful scale

[PlanckLimit81]

Fair correction. The algorithm gap is domain specific. Molecular simulation has it better than optimisation which has it better than general purpose quantum advantage

[DotEXE]

The Willow benchmark story is the perfect example of what you are describing. The classical computer would take longer than the age of the universe framing is technically accurate. The problem is the task is useless outside of benchmarking

[EdgeRatedR]

I would push back slightly. Below threshold error correction, which Willow demonstrated, is not a useless result. It is the foundational condition for fault tolerant QC. The problem is how it was communicated, not what was achieved

[ProperMadlad20]

Agree with that distinction. The achievement is real. The way it was written up in every tech outlet was misleading because it implied useful quantum advantage rather than a proof of concept for a key mechanism

[SerialScroller]

The 1800 to 2200 error correction specialists globally is the number that people should be sitting with. That is smaller than the number of people who worked on the Manhattan Project

[GhostRider89]

It is a tiny community for something this consequential. And they are being asked to solve what is essentially one of the hardest open engineering problems in history

[MickFoley00]

On the 2029 to 2032 timeline for fault tolerance. I think that is the optimistic read of the IBM and Google roadmaps, not an independent estimate. Actual delivery on hardware roadmaps in this field has historically lagged by years

[TheGreatMoney]

IBM's own roadmap has slipped before. The 2025 milestones from their 2022 roadmap were not all met. I would add at least two years to any announced timeline as a prior

[MondayMoan31]

The China situation is something nobody in the Western tech press covers properly. Chinese quantum cryptography deployment is real and happening now. The commercial race is not only about gate based quantum computers

[TommyB_20]

The QKD network stuff is interesting but it is a different technology to gate based QC. Conflating them is another version of the press release problem

[Ellie22]

True. Quantum key distribution and universal quantum computing are not the same thing and have different timelines and use cases. The press tends to smoosh them together

[NorthernKernel]

What is the current honest state on neutral atom architectures relative to superconducting. The ETH Zurich work this month suggests neutral atoms are catching up faster than people expected

[CrimsonFury]

Neutral atoms have real advantages in connectivity and coherence time. The ETH swap gate paper is interesting. The practical question is whether you can do it reliably enough and fast enough at scale, which is still open

[VB]

The 2040s timeline for broad commercial impact is the one I never see in mainstream coverage. Everyone writes about quantum as if the relevant horizon is 2028 at most

[DiamondDallas86]

Because 2040 does not drive clicks or investment rounds. The honest timeline is commercially inconvenient for everyone involved in funding or covering the field

[Dom9]

The show-me phase reflected in flat budgets is actually healthy. It means the field is being judged on results rather than promises. Painful short term and good long term