The future of computing runs on silicon-based quantum computing processors
Contrary to popular belief, simply demonstrating quantum advantage is not the ultimate goal of quantum computing. It would be an amazing feat of science and engineering, but it is not sufficient to reach the goal of making quantum computing mainstream. As we approach the era of quantum advantage, demonstrated by the continuously increasing capabilities of quantum computers around the world, the ecosystem may pivot its focus from scientific experiments and demonstrations to prioritizing scalability in the real world.
By Himadri Majumdar, Chief Executive Officer and Co-founder of SemiQon
The race to quantum supremacy will largely be solved by having scalable, reliable, and affordably manufacturable qubits that can get us to the million-qubit era to solve real-world problems.
Significant progress in the quantum computing communities all around the globe is already happening. After the first successful demonstrations of quantum computers in action, the next step is to integrate quantum computers or processors with high-performance computing (HPC) environments – supercomputers. This integration will demonstrate the benefits of coupling quantum computing with classical computing, which will be followed by further advancement in quantum computing capacity, all leading to the era of scalability in quantum computing.
But we are not there yet. Right now, we are at the crossroads where the global quantum computing ecosystem is primarily engaged in highly costly scientific experiments that focus on building stable qubits within the physical limits of superconductor, ion-trap, neutral atom or photonic technology – or identifying a new path.
The former keeps the status quo while the latter allows us to approach the next stages of quantum computing as difficult yet solvable engineering challenges. Most of the mentioned modalities are also working on the scalability aspects, even though aspects of practical and optimal module sizes, qubit connectivity and interconnect compatibility are still a work in progress.
The first-generation quantum computers have already convincingly demonstrated that harnessing the power of quantum phenomena for computation purposes is possible. The real challenge with further progress, if we can call it an era of practically useful quantum computing – or Quantum Computing 2.0 – is how it can be done in a scalable, sustainable, and affordable way.
Answers to the problems hindering the scalability and widespread adoption of quantum computing can be found much closer to home than most would think. To address the challenges of sustainability, scalability, and affordability, we must put the vast knowledge of semiconductor manufacturing to work to drive the future growth of the quantum computing industry.
Unlocking the full abilities of quantum computers can, at the moment, seem like a distant dream with many challenges, both from a hardware and software perspective. However, the history of technology and computing has taught us that what at one point in time seems almost impossible becomes a daily reality, even mundane, in the future.
Take, for instance, mobile phones. What began as a big, bulky machine has now turned into an incredibly powerful mini computer with mind-boggling computational ability that most of us have with us at all times – even young children. In fact, the astronauts who went to the Moon had far more primitive technology helping them navigate space and safeguard their lives.
So, what needed to happen for that phone that you use to send email, play games, and take photos with, to appear in your hand? It wouldn’t exist without the invention of semiconductor (silicon) integrated circuits or IC chips that paved the way for computation as we experience it today on our phones, tablets, smart TVs, and home devices. We certainly wouldn’t have AI – once a sci-fi concept that has now turned into a household topic – machine learning, metaverse, or the digitalization of many services, like healthcare and mail, which we take very much for granted.
In the 1940s, early computers such as ENIAC took up a large room and a lot of effort to run. With hundreds of wires and, sometimes, vacuum tubes connecting different parts of the machine that had to be manually adjusted, ENIAC weighed about 30 tons and was almost 2.5 meters tall and 30 meters in length. Additionally, it consumed a huge amount of electricity to complete computational tasks – 174 kW to be exact.
Surely in the 21st century, quantum computers are much more sophisticated than this? You’d be tempted to think so, but the facilities a functional quantum computer requires are even more complex than ENIAC. The chandelier, as the quantum computer is often fondly called due to its appearance, has hundreds of wires, and the more qubits it has, the more wires it needs. They require specific, low temperatures to function, several millions in financial investment, and a highly qualified team of quantum physicists and computer scientists to work on them and run them.
What this all means is that like ENIAC, modern quantum computers are simply not scalable and we need to look to the future to a moment when quantum computers finally begin to scale and deliver on their promise to develop vaccines, new materials and solve the practical questions of more far-off ideas like space travel.
But how do we get there and bring the promise of quantum computing down from its ivory tower to a productized level and real-life applications? Instead of relying on non-standard, exotic material-focused approaches to build quantum machines, it is possible to use proven and reliable semiconductor manufacturing technology for building quantum computing processors.
SemiQon is already on the path to demonstrating this. SemiQon builds its hardware using a customized CMOS fabrication process. In it, SemiQon’s team monolithically integrates silicon quantum dots and cryogenic CMOS on the same chip to build the full quantum processing unit or QPU. The benefit of this approach is that it allows the fabrication of significantly smaller qubit dimensions than what is typically used in current quantum computers, making it possible to pack more qubits into a smaller footprint. This in turn reduces the need for large-scale cryogenics or cooling to perform computations at the low temperatures standard quantum computers require.
The requirement of extremely low temperature can also be solved: we aim to operate our silicon-based quantum chips at higher temperatures than the norm – at 1–4 Kelvin – which means that less energy is needed, helping us achieve scalability and cost savings at the same time. Development is ongoing for on-chip cooling solutions that can potentially allow such temperatures to be achieved without significant infrastructure, too.
We shouldn’t stop there, however. We envision an era for a convenient plug-and-play processor that will allow integration of the processors into a highly scalable, full-stack quantum computer. We are on our way – we have our own fabrication operation at a national pilot line foundry and our own testing and measurement facility. We have now manufactured our first batch of processors and they are in use with our partners who are conducting different levels of testing around the world.
When it comes to the scalability of quantum computers, traditional quantum computing platforms are certainly ahead of the semiconductor-based platform. However, technological challenges are expected to bring the traditional approach to a plateau. The semiconductor quantum computing platform may have had a slower start, but once the scalability of manufacturing of semiconductor-based quantum processors is demonstrated, the approach can outpace all other platforms.
Scalability depends on several different aspects of the full stack of a quantum computer. This has been highlighted in various market reports and industry trends. The factors that come up often, and should be considered by all modalities are usually identified as fidelity at scale, control over individual qubits at scale, computation speed, multi-qubit networking, cooling requirements, and manufacturability – as discussed by McKinsey and Global Quantum Intelligence.
Predicting the impact of emerging and novel technologies is always an exercise for the imagination. However, we can undoubtedly expect that the million-qubit era of quantum computing will expand our idea of what is and is not possible. As we discuss the future, it is valuable to note that developments in quantum do not exist in a vacuum, separate from other scientific discoveries. For example, during the last century, humanity witnessed exponential growth in computing power walk – and run – hand in hand with other discoveries, such as space travel and the onset of bringing AI into our daily lives.
New milestones achieved by the quantum computing ecosystem will encourage businesses, research organizations, and governments to venture deeper into the world of quantum computing and to partner with companies that are paving the way to the million-qubit era.
Exploring the opportunities of this previously unattainable computing power should not be left to quantum scientists alone. The quantum ecosystem will greatly benefit from engaging existing industries, potential end-users, and interested policy-makers to help build technologies that solve problems, add value, and are easy to adopt. Realizing the promise of quantum requires strategic investments in both innovations and the existing semiconductor industry. The good news is that success in scaling up quantum will provide endless opportunities for industry and a competitive edge on the global
scale.
(This article is based on the presentation and the paper “Future of Computing: Silicon-based Quantum Computing Processors”).
