A Chinese quantum startup just launched a cloud platform with 100 qubit chips and a roadmap all the way to 2030

[Reacher Mitchell]

A new option for anyone wanting hands on access to real quantum hardware

Logical Qubit Technology hosted its Quantum Day 2026 event on July 9, officially unveiling a superconducting quantum computing cloud platform aimed at universities, research institutions, and commercial users across biomedicine, chemistry, materials, energy, AI, finance and logistics. The pitch is straightforward, lower the barrier to accessing genuinely capable quantum hardware, since full quantum computer systems remain prohibitively expensive and useful quantum cloud capacity has stayed scarce

Three levels of access, from simple to laboratory grade

The platform offers access at three distinct levels. At the physical qubit level, users can run arbitrary single qubit gates in 20 nanoseconds with up to 99.97 percent fidelity, submitting algorithms as simple code that gets translated into quantum circuits automatically. At the logical qubit level, built around a newer AGate-165 chip, readout time has been cut from 2 microseconds down to about 600 nanoseconds, with distance-9 repetition code and distance-3 surface code algorithms available for researchers to test error correction schemes directly. At the pulse control level, advanced users can directly manipulate the physical waveforms controlling each qubit with 1 nanosecond timing precision, effectively running a full quantum lab remotely without owning any hardware themselves

Real partnerships, not just a chip launch

The company announced a strategic partnership with pharmaceutical firm Pharmaron specifically to explore how quantum cloud platforms and drug companies can collaborate through joint laboratories, alongside partnerships with several quantum software companies building on top of the platform. CEO Zhen Wang also shared an useful data point on the field's overall progress, coherence time in superconducting qubits has improved three orders of magnitude industry wide over the past two decades, from 1 microsecond to the millisecond range, and Logical Qubit Technology itself recently hit over 890 microseconds on a small scale chip, with plans to transfer that improvement to full hundred qubit systems next

An unusually specific public roadmap

The company laid out year by year targets through the end of the decade, aiming to break through the fault tolerance threshold with error rates around one in a thousand by 2027, achieve deep integration with drug discovery and materials simulation by 2028 that actually beats classical computers on real problems, and complete engineering of tens of thousands of physical qubits with error rates down to one in a hundred million by 2029. Whether that timeline holds is obviously an open question, but publishing this level of specific, falsifiable commitment is an useful contrast to the vaguer roadmaps most quantum companies tend to offer

Logical Qubit Technology Launches Quantum Cloud Platform for Superconducting Quantum Computingthequantuminsider.com

[Mike]

Publishing a specific year by year roadmap with actual falsifiable numbers instead of vague someday promises deserves real credit, it's easy to check back against this in a couple years and hold them accountable

[RVD17]

Being able to directly manipulate pulse level waveforms with nanosecond precision remotely is genuinely laboratory grade access, that's a lot more than most cloud quantum platforms offer casual users

[Nathan75]

Coherence time improving three orders of magnitude industry wide over two decades really puts the pace of this field's progress into perspective when you step back and look at the long arc

[RainyDayFund]

Cutting readout time from 2 microseconds to 600 nanoseconds sounds like a small number but that kind of speed improvement compounds significantly once you're running millions of operations

[Woven Sasha]

The Pharmaron partnership for joint drug discovery labs feels like exactly the kind of real world commercial validation this field needs more of instead of just chip specs and benchmark claims

[Orbit William]

2027 for breaking the fault tolerance threshold is an aggressive target, curious how this compares to the timelines other major quantum companies are currently projecting for the same milestone

[Connor97]

The naming explanation, A Gate meaning a door, is a nice bit of branding that actually ties back to something meaningful about what they're trying to build here

[TinyGriffin]

That 2027 fault tolerance target feels... ambitious in the way quantum roadmaps often are

Hitting fault tolerance is not just about qubit count, it is about error rates, coherence times, and error correction overhead all lining up at once.

A 100 qubit chip sounds impressive on paper, but if the physical error rate is still too high, you end up needing thousands of physical qubits per logical qubit.

That is where most roadmaps quietly stretch.

Still, having cloud access is a big deal.

Even noisy systems are useful for testing algorithms and training people.

So even if the timeline slips, the platform itself could be valuable.

[RayOfLight32]

Feels like every few months there is another "we will reach fault tolerance soon" announcement

Not saying this one is wrong, just that the history of quantum timelines is... optimistic.

Compared to IBM and Google, the gap is usually in published error rates and reproducibility.

If this startup can show competitive fidelity numbers, then people will start taking the roadmap more seriously.

Until then, it is a bit of wait and see.

That said, more players in the space is great.

Competition tends to push progress faster.

[Darren51]

The cloud angle is actually what caught my attention more than the qubit number.

Making hardware accessible lowers the barrier for researchers and developers who do not have lab access.

That is how ecosystems grow.

Think about how classical cloud computing exploded once access became easy.

Quantum might follow a similar path, just slower and weirder

Even limited machines can be useful for hybrid algorithms.

So this could quietly matter more than the headline specs.

[Myles]

Comparing timelines, IBM has been talking about error-corrected systems closer to the end of the decade, while Google has hinted at earlier demonstrations under very controlled conditions.

So a 2027 claim is aggressive but not completely out of nowhere.

The difference is usually in what counts as "breaking" the threshold.

Lab demo versus practical system are very different things.

That nuance tends to get lost in announcements.

Still, if they even get close, that would be a serious achievement

Worth watching how they define success.

[Olivia78]

Part of me wonders how much of this is technical versus strategic signaling.

Announcing a long roadmap to 2030 sends a message about confidence and staying power.

Especially in a field where funding and perception matter a lot.

It is not just about the hardware, it is about attracting talent and partnerships.

That said, delivering even half of that roadmap would be impressive.

Quantum progress tends to be nonlinear.

Sudden jumps do happen, just not always on schedule

[Lantern76]

One thing that does not get enough attention is software tooling around these platforms.

If their cloud offering includes good SDKs, simulators, and documentation, that could make a huge difference.

Bad tooling can make even good hardware painful to use.

Good tooling can make mediocre hardware useful.

That balance matters more than people think.

Hopefully they learned from what IBM and others have done right.

Otherwise it becomes another "cool but inaccessible" system.