On the Verge of a Quantum Leap?

For over 60 years, developments in computer science have led to spectacular advances: high-performance computing has now reached exascale (1 billion billion FLOPS), and intelligent machines of a new kind are beginning to emerge...

Today, one of the main challenges for these technologies is energy efficiency.

Frontier at the Oak Ridge National Laboratory, officially the fastest computer on Earth, and one of the most energy-efficient in its class, needs a whopping 22.7 MW of electrical power to operate normally. 

Zettascale computing, the industry's new 'North Star', which will run 1,000 times faster than Frontier, promises exciting new horizons but will likely require dedicated nuclear reactors to process information: hardly a sustainable option, then…

There are of course a number of possible avenues for improvement, all of which are already  attracting  meaningful  private  funding – innovative cooling systems, 3D-microchip  architectures, and photonic transport to name a few. But none of this guarantees a financially viable outcome.

Rather, a complete change of  paradigm  might  be  needed.

Enter Quantum computing, a largely misunderstood technology that harnesses the laws of quantum mechanics to solve problems that are too complex for classical systems.

Here, we seek to highlight 5 key opportunities and dispel 5 common misconceptions that could prevent investment in this developing field.

At least half a dozen different approaches are competing, with no clear consensus on their chances of success, but we are particularly impressed by quantum photonics, which builds on the fabrication techniques of the semiconductor industry and provides robust microchips that can operate at room temperature and interface with conventional devices, away from constraining cables, dilution refrigerators or electromagnetic traps. Such scalable hardware has the potential to transform the way we approach complex problems and to drive innovation and growth in a wide range of verticals.

The companies in question focus on hardware development and manufacturing, algorithms and user-friendly software solutions, quantum error mitigation and correction, but also training and education. They understand that their success depends on all these pillars at the same time, and they have become very effective at creating entire environments and communities around their main quantum computing products. 

Because of the strategic nature of the technology, governments will continue to support quantum innovation projects. Public-private partnerships will multiply and help stabilise the market in the long run, as is already happening in other areas of deep physics, such as nuclear fusion.

Although we lack the data to formally support our position and it is impossible to draw a definitive conclusion until large quantum computing platforms are operational, Google and other companies have suggested that the environmental impact of quantum technologies will be significantly less than that of their conventional equivalents.

The field of quantum metrology is highly promising. Quantum sensors can detect infinitesimal changes in magnetic fields to map neural activity and advance our understanding of the human brain; they can be used to accurately measure gravitational fields with applications in navigation and geophysics; they can help detect the minute presence of chemical compounds for environmental monitoring, medical diagnostics, cancer screening, or food safety...Secure Quantum communications is another domain of interest – in particular, Quantum Key Distribution and Quantum Key Distillation techniques that leverage photon properties to secure classically encrypted messages.

Quantum computing is sometimes ascribed “quasi-magical” properties but while the concept is not easily translated from algebraic notation to plain language, it is only about creating linear combinations of states, each weighed by complex coefficients describing their amplitude and relative phases and interfering constructively  or destructively   with   each   other   to   define   a   probability distribution.  When  this  is  done  intelligently  through  a  well-crafted  algorithm, the  correct  solution  to  a  problem  is  extracted  with  a sufficiently high probability of success. Clearly, no magic here.  

Quantum computers will be able to tackle certain problems involving the manipulation of large numbers exponentially  faster  than  their  classical peers, but  this  will  always  depend  on the mathematical structure of the problem. In many situations, a conventional computer will be just  as  suitable – if  not  more  suitable – than  a quantum processor. It follows that the identification of the problems for which quantum mechanics  can  provide  a  real  advantage  and  the  design  of  the  necessary algorithms to address them is crucial to the development of the industry. So far, just a few useful algorithms have been proposed. Quantum computers will likely be part of larger hybrid structures, alongside classical and AI-powered processing units, to be tasked selectively as part of a wider toolkit.

Certain analogue quantum computers are ready to tackle specific optimization problems (in logistics, finance etc...) although it is unclear how much of an advantage they really provide at this stage. Meanwhile, many of the newest quantum algorithms for business applications currently run on simulators, i.e., in a classical framework with a small number of perfect virtual qubits. Some computational chemistry is possible, as in determining the ground-state energy of simple molecular  systems, but larger quantum systems will be needed to achieve truly revolutionary breakthroughs on an industrial scale. Although a growing number of companies are willing to explore the possibilities offered by quantum computing, notably through partnerships and collaborations with major hardware suppliers, adoption is likely to remain modest until universal quantum computing becomes a reality.

Despite powerful advocates of the concept, we remain sceptical. Companies like IBM firmly believe that noisy (error-prone) physical systems of a few hundred qubits in a robust error suppression and mitigation environment can help accelerate progress in various scientific domains. But after four years of development, NISQ machines have yet to find their place and still seem a long way from a viable commercial use case. Without burying them completely, the pursuit of Full-Scale Fault Tolerance could be a more valid objective, despite the immense challenges in its path.

Significant progress has been made, but much more work is required for surface codes (or any other approach) to detect and correct errors in a scalable, resource-efficient way. Talks of systems of a million qubits do not inspire confidence, when the current state of the art barely exceeds one thousand. In all likelihood, quantum error correction may not be viable for a few years yet.   

For the time being, quantum computing remains in a "superposition" between greatness and obscurity. However, despite the scale of the technical challenges ahead, there are grounds for optimism and the potential benefits remain considerable. For those prepared to take a long-term view of the sector's development, the rewards could be substantial.