Quantum Motion Claims They Built the First Silicon CMOS Quantum Computer

This Could Actually Matter for Scaling

Quantum Motion just deployed what they're calling the world's first silicon CMOS quantum computer at the UK National Quantum Computing Centre. The interesting part: they built it using regular chip fabs instead of the exotic custom shit every other quantum computer needs.

It fits in three server racks, which beats IBM's building-sized superconducting systems. Whether it actually works well enough to matter is a totally different question.

Why Silicon Manufacturing Could Be Huge

Every other quantum computer is basically a one-off prototype. IBM and Google's superconducting systems need custom dilution refrigerators that cost more than your house. IonQ's trapped ion systems require ultra-high vacuum chambers and laser precision that breaks if you look at it wrong. These approaches work in the lab but they're fucking impossible to scale or manufacture at any reasonable cost.

Quantum Motion's approach uses spin qubits on standard 300mm silicon wafers - the same manufacturing process that cranks out every CPU and GPU. CEO James Palles-Dimmock claims this means they can "mass-produce" quantum computers using existing fab infrastructure.

That sounds great on paper. In practice, I've seen too many quantum demos that fall apart under real testing. Until they publish actual coherence times, this is just expensive silicon.

The Manufacturing Reality Check

The semiconductor industry has burned through $2 trillion over 50 years getting silicon CMOS manufacturing to actually work. Every major fab, supply chain, and quality control process is built around silicon. If Quantum Motion's approach actually works, they could leverage all of that instead of starting from zero with custom quantum manufacturing.

Look at the alternatives:

• Superconducting qubits need dilution refrigerators at 0.01K (good luck with that)
• Trapped ions need ultra-high vacuum and laser precision that drifts if someone sneezes
• Neutral atoms need complex optical lattices that take months to align

Each needs specialized facilities, exotic materials, and teams of PhDs babysitting the damn things 24/7. Quantum Motion uses the same foundries that shit out billions of regular processors annually.

The Scaling Promise vs Reality

Ensar Seker from SOCRadar pointed out the obvious advantage: "Silicon-based quantum architectures leverage decades of investment in CMOS manufacturing, supply chain maturity, and quality control. This contrasts with other technologies like superconducting qubits or trapped ions, which typically require highly specialized environments."

Quantum Motion claims their approach can scale to "millions of qubits per QPU" because they can tile the design across a chip. That's the theory. In practice, quantum error rates usually get worse as you add more qubits, not better. I learned this the hard way working on quantum error correction - more qubits means more decoherence, more crosstalk, more everything that breaks.

Breaking RSA-2048 needs about 4,000 error-corrected qubits. Current quantum computers have maybe a few hundred noisy ones that work for microseconds. The gap between "millions of qubits" and "4,000 useful qubits" is still fucking enormous.

Industry Skepticism and Reality

Sam Lucero, a quantum strategy consultant, acknowledged this is "the first full implementation of a silicon spin qubit computer" he's aware of, but added the important caveat: "Since there is no performance data, it's not clear how this machine will compare to other available platforms at the moment, but I'd expect it to be fairly rudimentary in comparison."

Translation: it probably doesn't work very well yet. When quantum companies won't publish benchmarks, that usually means the numbers suck.

Professor Prineha Narang from UCLA was more optimistic, noting that "solid-state quantum technologies are catching up to the superconducting and atomic platforms."

The Encryption Timeline Question

Every quantum computing announcement comes with the same encryption panic. DigiCert's Tim Hollebeek gave the standard scary line: "Quantum computers are scaling up rapidly... Advances like Quantum Motion's continue to bring us one step closer to their eventual existence."

The industry consensus still puts cryptographically relevant quantum computers around 2029. Quantum Motion's silicon approach might accelerate that timeline if it actually scales, but there's a massive difference between building qubits and building qubits that don't lose coherence in 10 microseconds.

What This Actually Means

Quantum Motion's achievement is significant because they proved you can build quantum computers using standard chip manufacturing. That's genuinely important for eventual scaling and cost reduction.

But they haven't proven their qubits work well enough to beat a goddamn calculator at anything useful. No performance data, no benchmarks, no comparisons. Just the fact that they built something in three server racks using regular fabs.

The real test comes when they publish actual performance numbers. Can their silicon qubits maintain coherence for longer than 50 microseconds? How do error rates compare to IBM's 0.1% gate errors? How much does manufacturing cost actually drop? These are the questions that matter.

Hugo Saleh, Quantum Motion's President, claims they're "on track to bring commercially useful quantum computers to market this decade." Every quantum company says this shit. I've heard identical claims from D-Wave, Rigetti, IonQ, and everyone else. The difference is whether their approach can actually scale beyond a demo in a lab.

Building quantum computers in standard fabs is clever. Making them work reliably enough to matter is the hard part. The error rates matter infinitely more than the marketing hype.

Why Silicon Success Doesn't Guarantee Quantum Success

Quantum Motion's silicon approach solves the manufacturing problem, but manufacturing was never the hard part of quantum computing. The hard part is making qubits that actually work reliably enough to run useful algorithms.

Silicon spin qubits have been a research area for over a decade. Intel's been working on them since 2015 with their QuTech partnership. The challenge isn't building them - it's controlling them precisely enough to perform quantum operations without errors destroying the computation.

Error Rates Are Everything

Current quantum computers fail because quantum states are fragile. Any interaction with the environment destroys the quantum information through decoherence. IBM's best superconducting qubits have error rates around 0.1% per operation. For useful quantum algorithms, you need error rates below 0.01%.

Quantum Motion hasn't published error rates for their silicon qubits. That's the most important number missing from their announcement. You can manufacture millions of qubits, but if each one has a 1% error rate, they're useless for anything beyond demos.

The Scaling Challenge

Quantum error correction requires hundreds of physical qubits to create one "logical" qubit that's actually useful for computation. Current quantum computers are in the NISQ era - Noisy Intermediate-Scale Quantum - because they don't have enough good qubits for error correction.

Breaking RSA encryption needs about 4,000 logical qubits, which translates to millions of physical qubits if error rates don't improve dramatically. Quantum Motion's claim about "millions of qubits per QPU" sounds impressive until you realize that's what you need just to get started on useful problems.

What Success Actually Looks Like

If Quantum Motion's approach works, it won't be because silicon manufacturing is better. It'll be because they solved the fundamental physics problems of qubit control, error rates, and quantum coherence. The manufacturing advantage only matters if the qubits actually work.

The real test comes when they publish benchmarks comparing their silicon qubits to IBM's superconducting systems or IonQ's trapped ions. Can they maintain quantum coherence long enough to run useful algorithms? How do their error rates compare? How long can they maintain quantum states before decoherence destroys the computation?

Building quantum computers in server racks is clever marketing. Making them work reliably enough to solve problems classical computers can't handle is the actual challenge.