Diraq Produces Quantum Chip Using Standard Semiconductor Process

ICM Limited's investee Diraq recently announced that its quantum processor designs have been successfully manufactured using industry-standard materials and processes. This is a major proof point for building full-scale quantum computers on single chips using the same processes that make the chips that power our phones.

The qubit (or ‘quantum bit’ — the building block of a quantum computer) produced at imec is the most high-quality qubit demonstrated to date for a silicon spin qubit made in a semiconductor foundry environment. The result highlights Diraq’s key advantage of using existing semiconductor processes. Unsurprisingly, it turns out you can make better qubits with billion-dollar machines. Consider that there are multiple rounds of design optimisation of chips in semiconductor manufacturing so to have a high-quality result as a first demonstration is a tremendous achievement, and we anticipate Diraq will be able to push the quality of their qubits much higher as this partnership continues. This team has already achieved world-record quality qubits using the crude tools of a university lab and we expect its foundry devices to eventually exceed those achievements.

Another key part of this release is that Diraq’s designs will be manufactured this month at GlobalFoundries (‘GF’), one of the world’s most important semiconductor manufacturers. The ability of the semiconductor industry to provide exponentially more computing power at lower cost has underpinned the digital age. Diraq is therefore taking advantage of 60 years of technology development and trillions of dollars of investment to produce devices at scale. Sarah Constantin’s simplification in The Enchippening is that this advantage means “you should be bullish on any useful tech that is newly fabricable on a chip”.

It’s not the first time a quantum computing company will produce a chip at GlobalFoundries, with both PsiQuantum and Xanadu announcing partnerships just months before reaching unicorn status. While PsiQuantum and Xanadu use GF Fotonix, GF’s silicon photonics platform, Diraq will use the FDX FD-SOI platform, which makes transistors like those found in your phone at the scale of billions. A key difference is that spin qubits in FD-SOI are the same size as transistors, which are around a million times smaller in area than the photonic waveguide qubits produced by Fotonix. This means you can pack many more qubits into one chip, potentially billions. It’s also worth noting that FD-SOI is a much more mature technology than Fotonix which brings a range of benefits including widespread expertise, understanding of limitations and opportunities, and predictability.

Building a quantum computer is already hard and we know that it will take a large number of high-quality qubits — likely many millions of physical qubits — to deliver on revolutionary use-cases like in chemistry or drug-discovery but for a company to be commercially viable, it must also be able to produce quantum computers at scale and provide access to them at reasonable cost. Utilising existing, world-leading infrastructure will produce consistent, high-quality devices and is much more cost effective than building your own factory. These announcements should raise some critical questions in the industry about the way different types of quantum computers can be manufactured at scale. Diraq, by using essentially modified transistors (this type of qubit is often referred to as “spins in silicon”), have shown a pathway to manufacturability that does not require hundreds of millions of dollars of capital expenditure.

The size of the quantum computer produced is also important, and spins in silicon is the only modality with a theoretical basis to hold billions of qubits on a single chip. This chip will need to be cooled down (to Space-level temperatures), meaning it’s unlikely we’ll ever have a quantum chip in our phones, but a quantum computer in a single unit around the size of a small fridge makes a lot more commercial sense than one the size of a football-field. Part of the problem is the unit economics of such a machine, but physical footprint, power consumption, carbon emissions and portability also make it difficult to believe in the commercial viability of proposed devices. We are looking forward to sharing further analysis on this topic.

Putting these two ideas together — manufacturing at scale and small physical size — means Diraq will benefit from economies of scale, making it difficult for others to compete on price.

In our view, spins in silicon are the only pathway to a commercially useful and viable quantum computer. There’s a still a long way to go and engineering to do but we back Diraq, the inventors of this technology, as the team to deliver it.

You can read the press release here.