IBM said it plans to launch quantum computers with 4,000 qubit processors by 2025 and put quantum computing into practical use. The company shared its roadmap for building a modular architecture that will enable quantum computers to be further scaled. In addition to increasing the number of qubits, the company is also working on creating software to “distribute workloads across quantum and classical resources.”
Qubit or quantum bit is the basic unit of quantum information. Higher qubits mean more computing power, which translates into more powerful quantum computers.
So far, IBM has released quantum processors of up to 127 qubits. Last November, IBM said it was working on two new processors with a lot of computing power. The processor named Osprey, which will have 433 qubits, is expected to be launched this year in 2022, while another processor called Condor with 1,121 qubits is expected to be launched in late 2023.
With the modular architecture, IBM is hoping to use these processors to achieve the equivalent of a 4,000 qubit processor. IBM said it would deploy chip-level couplers that would link multiple chips together to effectively make a single large processor.
Jay Gambetta, Vice President of Quantum Computing, said, “By combining a modular quantum processor with a classical infrastructure orchestrated by the CuSkit runtime, we are creating a platform that allows users to easily build quantum computations into their workflow and meet the pressing challenges of our times.” will help deal with it.” and IBM Fellows said in a statement.
Qiskit Runtime is a containerized runtime software that runs on the IBM cloud and uses classical computers to optimize workloads and then execute them efficiently on large-scale quantum systems. Other companies like Fujitsu are also looking for something along the same lines. Last month, the Japanese company unveiled its quantum computing simulator that will run on the company’s supercomputers and run quantum programs to test how qubits will respond to various operations when run on a real quantum system.
In addition, IBM said it would develop software to improve error suppression and mitigation in quantum computers. Unlike classical computers, the laws of quantum physics restrict how error correction works in quantum computers. This makes them highly prone to interference, leading to errors in the quantum algorithms that run on it.
Many researchers in the field are working to solve this problem. In January, researchers from three universities in different countries demonstrated how the traditional silicon used in chips could solve the high error rates in quantum computers.
IBM has been one of the forerunners in quantum computing and in 2019 launched its first quantum system with 20 qubits for commercial use.
India has also partnered with countries that have made early inroads into quantum computing. Last month, Finland signed a joint declaration with India to set up an Indo-Finnish virtual network center on quantum computing called QNxT.