In the following, I will give you a detailed overview of the quantum computer offerings that already exist and what concrete plans the industry is pursuing for the future. You will gain an insight into what a quantum computer is and how it can be used. I’ll update this article regularly and keep you up to date with the latest developments in the quantum computing industry.
When I first wrote this article in 2018, the situation of commercial hardware manufacturers of quantum computers and their architecture strategies was pretty clear. After adapting it a few times in the meantime, almost three years later I find that I should rewrite the text in large parts. The key players from three years ago have established themselves further and are showing impressive results. But other companies have since appeared on the scene, stirring up the market. The same applies to the architecture strategies: while in 2018 it still seemed as if there would be no other serious quantum computer technology apart from superconducting qubits in the foreseeable future, the year 2020 was probably the year of the ion trap quantum computer. However, the annual financial statements belonged to the photonic quantum computers, which are picking up speed and whose manufacturers are currently collecting investor money with ease. This trend of diversification will probably continue in the future.
What is a quantum computer?
A quantum computer is a calculating machine that does not work with our everyday logic, but with quantum logic. The elementary computing units are quantum bits (“qubits”). Unlike conventional bits, qubits can assume any superposition of the “0” and “1” states. This is one of the fundamental properties of quantum mechanics (called “wave-particle duality”). I use a simplified qubit representation on quantum computer-info.de, which, however, reflects the essentials well.
For more information, check out my comprehensive guide The Incredible Quantum Computing Simply Explained. There you will learn even more exciting things about the principle of quantum computers. Among other things, I will use a vector image to explain how the “spooky” quantum entanglement works. You will also learn what a quantum computer program looks like and how you can quickly try one out yourself online. In addition, I will explain to you in a simple and very visual way how a well-known search algorithm for quantum computers (the “Grover algorithm”) roughly works.
In the network, the qubits and quantum circuits can interfere with each other (“crosstalk”).
What these sources of error and limitations mean in practice can be seen very clearly in the study “The Bitter Truth About Quantum Algorithms in the NISQ Era” by the University of Stuttgart v . Whether one will be able to achieve a quantum advantage in practical and industrially relevant applications with NISQ quantum computers is currently still open and the subject of intensive research vi. On the other hand, the number of companies in established markets that are planning budgets for quantum computing in the next few years is increasing rapidly.
To account for the limitations of NISQ quantum computers in a single metric, IBM’s Quantum team developed a protocol in 2018 to determine the “effective” number of qubits. The Quantum Volume metric is based on a benchmark protocol viii: To do this, one measures the largest possible “square” quantum program of a certain, reasonably meaningful type that delivers “convincing” results (“square” here means: the number of qubits used and the circuit depth should be the same).
When I first wrote this article in 2018, the situation of commercial hardware manufacturers of quantum computers and their architecture strategies was pretty clear. After adapting it a few times in the meantime, almost three years later I find that I should rewrite the text in large parts. The key players from three years ago have established themselves further and are showing impressive results. But other companies have since appeared on the scene, stirring up the market. The same applies to the architecture strategies: while in 2018 it still seemed as if there would be no other serious quantum computer technology apart from superconducting qubits in the foreseeable future, the year 2020 was probably the year of the ion trap quantum computer. However, the annual financial statements belonged to the photonic quantum computers, which are picking up speed and whose manufacturers are currently collecting investor money with ease. This trend of diversification will probably continue in the future.
Unlike conventional computers
A quantum computer doubles its performance potential with each additional qubit (i.e. an explosive or exponential increase). Current quantum computers are still small, with currently 10 to 65 qubits, and very limited (NISQ = “noisy intermediate-scale quantum”, more on this below in the text). A quantum computer with a size of about 26 perfect qubits can still be “simulated” with a normal laptop ii. However, the air for conventional computers then becomes thinner very quickly because of the rate of increase mentioned: From a magnitude of around 50 qubits, a quantum computer can no longer be stimulated by any current supercomputer (more on this in the paragraph on Google’s quantum computer) and after that, the differences become enormous. Due to its different logic, however, conventional programs cannot be transferred to a quantum computer, but completely new quantum algorithms are required for this, which are the subject of intensive research. Algorithms have already been found for some important applications that have a huge speed advantage over conventional methods (e.g. the “Shor algorithm” for prime factorization). However, they also share the fact that they require very large, error-correcting quantum computers.
For more information, check out my comprehensive guide The Incredible Quantum Computing Simply Explained. There you will learn even more exciting things about the principle of quantum computers. Among other things, I will use a vector image to explain how the “spooky” quantum entanglement works. You will also learn what a quantum computer program looks like and how you can quickly try one out yourself online. In addition, I will explain to you in a simple and very visual way how a well-known search algorithm for quantum computers (the “Grover algorithm”) roughly works.
Qubit Number, Quantum Volume, and Error Rates: The “Bitter Truth” About NISQ Quantum Computing
Until recently, the number of qubits was primarily used to classify the potential of concrete quantum computers. Here, superconducting quantum computers are currently clearly ahead. The metric makes sense since a quantum computer doubles its power with each additional qubit. At least on paper, because the current NISQ quantum computers have several crucial limitations:
Not all qubits can be interconnected in pairs (“connectivity”). For superconducting quantum computers, only the nearest neighbor qubits can usually be combined. This then requires additional, error-prone exchange operations (“swaps”). Here, too, ion trap quantum computers are much more flexible.
In the network, the qubits and quantum circuits can interfere with each other (“crosstalk”).
Additional errors can occur when initializing the qubits and reading the result (“readout error” or “SPAM error”, magnitude 1% or 1:100).
What these sources of error and limitations mean in practice can be seen very clearly in the study “The Bitter Truth About Quantum Algorithms in the NISQ Era” by the University of Stuttgart v . Whether one will be able to achieve a quantum advantage in practical and industrially relevant applications with NISQ quantum computers is currently still open and the subject of intensive research vi. On the other hand, the number of companies in established markets that are planning budgets for quantum computing in the next few years is increasing rapidly.
Limitations of NISQ
To account for the limitations of NISQ quantum computers in a single metric. IBM’s Quantum team developed a protocol in 2018 to determine the “effective” number of qubits. The Quantum Volume metric is based on a benchmark protocol viii: To do this, one measures the largest possible “square” quantum program of a certain. Reasonably meaningful type that delivers “convincing” results (“square” here means: the number of qubits used and the circuit depth should be the same). The quantum programs for the benchmarks must have a specific, random-based structure and check the restrictions mentioned above. However, for practical reasons, the quantum computer may first recompile this structure into a more hardware-friendly structure. An effective number of qubits “q” is determined from this square (the company IonQ also calls this number “algorithmic qubits”). The quantum volume is then the value 2 raised “q” to account for the exponential character of quantum computers.