Quantum computers: the hardware for AI
Constantly increasing requirements necessitate new hardware systems. The age of quantum computers is dawning.
Today’s computer hardware based on microelectronics is already reaching its technological limits. The era of huge leaps in development seems to have come to an end. For years now, people have been forecasting the end of Moore’s Law. The law postulated by Intel co-founder Gorden Moore in 1965 states that the performance of computer chips doubles on average every 18 months. This exponential growth in computing capacity is now reaching its physical limits. At the same time, requirements continue to rise, for example due to complex applications in the field of artificial intelligence. These requirements become even greater when combining different mega-technologies such as AI with nanotechnology, genetic engineering with cloud computing or big data with the Internet of Things. The resulting possibilities cannot even be estimated today and exactly the same applies to the complexity of such applications.
To make this possible, various companies have been working for years on a new generation of computers based on a different functional principle. Quantum mechanics is used instead of microelectronics. Put in a very simplified way, quantum computers no longer work exclusively with bits as the smallest memory unit, which can only assume the value 0 or 1, but also recognize a value in between, which are referred to as qubits. The binary system is thus extended by one dimension and can therefore represent information quite differently, since the smallest storage unit can now assume three values.
What do we need quantum computers for?
We are currently at the beginning of a comprehensive digital transformation of society as a whole, the exact extent of which no one can reliably predict yet. Already today, gigantic mountains of data are emerging from an increasing number of sources and the degree of networking will rise significantly. This data includes a plethora of useful information that can tremendously accelerate the progress of mankind. The digitalization and networking of all the world’s weather stations has made it possible, for example, to simulate weather and climate models that, thanks to artificial intelligence, eclipse everything we know today in this field. It would not allow the weather to be controlled, but make forecasting of extreme weather events much more accurate. This not only increases safety, but also allows concrete conclusions to be drawn about human impact on the climate.
Quantum computers are especially required to deal with large amounts of data and complex relationships. This is currently the case above all in research and science, but with the increasing development of mega-technologies it is also becoming more interesting for business (e.g. in the area of mobility).
For example, in its search for new business models, Volkswagen is experimenting with a new type of traffic management system. Anonymous motion data from transmitters in travelers’ vehicles and smartphones are analyzed to map traffic and passenger volumes. While this step can still be performed by conventional computers, optimization of the management system is then performed by a quantum algorithm. This precisely calculates the number of vehicles required in advance so that all road users can reach their destinations without waiting times and all vehicles can be utilized to capacity. The next increase in complexity would then be a completely automated traffic system.
Similarly complex scenarios, such as decentralized energy supply or smart city projects, may not only benefit from the capacity of quantum computers, but in some cases only be made possible at all as a result. In addition to artificial intelligence as the basic technology of the future, the requirements of real time and combinatorics will particularly necessitate fast high-volume computing. The energy transition from large megapower plants to many small decentralized energy producers not only requires smart distribution, but also precise forecasting of consumption and feed-in of renewable energies. If such an energy system is connected to the IoT on the one hand to analyze consumption and regulate it according to availability and, on the other hand, to the digitalized weather model to precisely calculate feed-in, then the computing requirements will supersede the capacity of today’s supercomputers.
The bottom line: quantum computers are an option, not a replacement
For many years to come, there will be many more tasks for conventional computers than for quantum computers. And it is not yet certain whether we will ever have a quantum computer on our own desk. However, we also know that in the initial days of microprocessors hardly anyone could have imagined we would someday all be carrying them around with us in our pockets. Nevertheless, the first commercially usable quantum computer, the IBM Q System, was presented at CES 2019. The first step has been taken and, although something similar to Moore’s Law will certainly apply here as well, further development will also be exponential. Then it won’t be long before we see the first applications for marketing, because data volumes and complexity are rising by leaps and bounds in this area as well and will soon require enormous computing capacities.