Unveiling Oxford Ionics’ development roadmap to scalable fault-tolerant quantum computing

Building quantum computers capable of realising quantum advantage is anything but easy. Dozens of companies are attempting it, pouring their expertise and resources into different strategies all with the aim of successfully bringing this technology to market. 

Why is it so hard? For starters, we’re going up against 70 years of progress and innovation in classical computing! To beat an excellent supercomputer, you need a good quantum computer. But on the flip side, the power of quantum computers scales significantly faster than classical computers – meaning that with just a little bit more progress, we get huge leaps in performance. 

In our very first blog post, we discussed how Oxford Ionics has taken a unique approach to this challenge by building classical chips that perform quantum operations. Through this pioneering engineering, we have put together the fundamental ingredients required to build the highest performing, most scalable quantum computers on the market. 

Today, we unveiled a development roadmap that outlines how we’re making this vision a reality. It includes three distinct short-term phases - Foundation, Enterprise-grade, and Value at scale - that take us to 10,000+ qubit platforms with a suite of capabilities that unlock valuable use cases for customers across all sectors. The systems delivered as part of the roadmap will power our longer-term technology strategy as we work towards delivering devices with 1M qubits and beyond. 

Let’s dive into the details behind our roadmap. 

First of all, why pay so much attention to fidelity? 

Reviewing our development roadmap, the first question you might ask is why are we anchoring so heavily on fidelity? From day one, we’ve held the philosophy that the quality of operations we run on quantum computers is just as important as the quantity. You might have hundreds or thousands of qubits today or tomorrow, but without high-fidelity qubits, the algorithms you can run are entirely limited. But…I can already hear your follow up question: isn’t that exactly what Quantum Error Correction (QEC) is for? 

Indeed, QEC is a fundamental tool to improve the quality of operations, such that one can reach error rates as low as 10^-6 (the MegaQuops regime) to 10^-12 (the TeraQuops regime). However, as with most things in life, QEC is not a free lunch. The number of high-quality qubits (defined as “logical qubits”) one ends up with after applying the QEC protocol depends incredibly strongly on the quality of the physical qubits one starts with. In other words, if the physical qubits are very noisy, it will take way more of them to end up with a handful of logical qubits. 

This is where high fidelity, or low error rate, physical operations really pay off. A single high-fidelity physical qubit is worth much more than hundreds of low-fidelity physical qubits. This allows us to approach physical to logical qubit ratios of 13:1 while achieving error rates as low as 10^-8, compared to ratios of 1000:1 or more required by other quantum computing platforms. Consequently, we can scale much faster, with less complexity, as we need far fewer physical qubits. In addition, thanks to our ultra low physical errors, we can engage a far wider range of sophisticated algorithms – this is because fewer errors enable longer and more complex circuits to run successfully before any QEC is required.

So for our development roadmap, we’ve decided to anchor it not just on qubit count, but in achieving the industry’s lowest error rates of of 10⁻⁴ across the largest number of physical qubits. 

Unpacking our development roadmap 

Foundation

In our Foundation phase, we are developing systems with 16-64 physical qubits at 99.99% fidelity. We have multiple systems based on the Foundation platform in operation at our headquarters, and are already deploying them onsite to customer environments. These platforms, based on our proprietary trapped-ion architecture, feature all-to-all connectivity between qubits and the ability to run operations in parallel. Both of these features are key to running efficient quantum algorithms. 

The Quantum Processing Units (QPUs) that power this platform contain all the elements that underpin our long-term development roadmap, including our proprietary all-integrated electronics capable of trapping and controlling our qubits.  Each QPU is composed of unit cells that can be tiled in a 2-dimensional grid – each able to host and control a small number of interacting qubits – making this platform easy to scale and highly flexible. The QPUs are made of multiple integrated layers, including RF and DC electrodes to drive global and individual quantum operations, and incorporating “junctions” to efficiently transport and entangle distant qubits in a Q-CCD architecture. 

The first key milestone for these systems is achieving one of the world’s highest Quantum Volume (QV) of 2^16. This performance will empower organisations to engage in research on quantum algorithms, QEC, and early applications to fields like security and defence, pharmaceuticals, and materials science

Enterprise-grade 

In Enterprise-grade, we target 256+ physical qubits at 99.99% fidelity. These systems, which can be ordered today, will be the most powerful quantum computers on the market. As we continue adding ultra-high fidelity qubits to our chips, we also expand their capabilities, introducing mid-circuit measurement and feed-forward – a necessary feature to run effective Quantum Error Correction (QEC) protocols, as well as sophisticated quantum algorithms, such as those required to solve computational fluid dynamics problems. Enterprise-Grade systems are QEC-ready, and will allow for over 16 logical qubits (with 10^-8 error rates). These capabilities enable advanced quantum applications development across industries such as materials design, aerospace, finance and pharmaceuticals. 

Building upon our technology stack, at the QPU level we increase the number of elements we integrate into the chip, including integrated photonic elements to cool and readout large numbers of qubits in parallel. While our QPUs get increasingly more complex to control larger numbers of qubits, they retain their essential simplicity which allows for rapid manufacturing through standard semiconductor foundries. And what’s more? We can upgrade a Foundation system to an Enterprise-Grade system by simply swapping the QPU.

Value at scale 

Our Value at scale phase builds off the work conducted in the Enterprise-grade period, enabling Oxford Ionics to scale its 256-qubit designs to 10,000+ qubits by replicating the design elements rather than reinventing. With this platform, we will be able to run QEC protocols yielding over 700 logical qubits at 10^-8 error rates. 

In order to achieve this, we will be developing an ultra-dense 2-dimensional fabric on a single chip. Each QPU will consist of a chip stack integrating all necessary photonics, passive and active electronics. In particular, we will be deploying our WISE (Wiring using Integrated Switching Electronics) architecture replacing large arrays of digital-analog converters (DACs) with far simpler multiplexers integrated in the QPU. Signal multiplexing greatly reduces the number of IO lines resulting in a single, all-integrated chip capable of hosting and controlling over 10,000 qubits. Yet the surrounding infrastructure will remain largely the same, simple and compact.

Whilst our Enterprise-grade systems will unlock early applications that surpass classical computing capabilities, our Value at Scale systems will realise broad commercial value – yielding solutions to previously-unsolvable problems across a wide range of use cases. This will truly be a new frontier in the field of quantum computing, and we’re incredibly excited to see the revolutionary impact these systems will have on some of the most challenging problems facing industries today. 

Delivering real-world value

Of course, it’s not just enough to deliver systems that only work in proof of concept demonstrations - they need to work in the real world, in real customer environments. We are already commercialising our quantum computers, and will be deploying machines to customers like the UK’s National Quantum Computing Centre and Germany’s Cyberagentur. 

But importantly, this is just the beginning. Our engineering teams are relentlessly focused on building the world’s most powerful quantum computers, capable of revolutionising how we solve complex challenges across all industries. And while we work towards realising this vision through building quantum computers with 1M+ qubits and beyond, the development roadmap we unveiled today will unlock ongoing value for our customers from day one. 

If you are interested in ordering one of the systems in our Foundation or Enterprise-grade phases, schedule a meeting with us by emailing sales@oxionics.com.