Technology company IBM announced progress in its quantum computing roadmap at its annual Quantum Developer Conference, including the upcoming release of its new processor and updates to its quantum software platform.
IBM introduced IBM Quantum Nighthawk, which it described as its most advanced quantum processor to date. The company said the processor is designed to work with high-performing quantum software and is expected to be available to IBM users by the end of 2025.
“There are many pillars to bringing truly useful quantum computing to the world,” said Jay Gambetta, director of IBM Research and IBM Fellow. “We believe that IBM is the only company that is positioned to rapidly invent and scale quantum software, hardware, fabrication, and error correction to unlock transformative applications. We are thrilled to announce many of these milestones today.”
The Nighthawk processor features 120 qubits connected through 218 next-generation tunable couplers, an increase of more than 20% compared with IBM Quantum Heron. This configuration will allow the processor to handle more complex quantum circuits while maintaining low error rates. IBM said this architecture could execute circuits with 30% greater complexity and tackle problems requiring up to 5,000 two-qubit gates, an important operation in quantum computing.
IBM expects future versions of Nighthawk to handle up to 7,500 gates by 2026, 10,000 gates by 2027, and 15,000 gates by 2028. These improvements will be supported by a larger network of qubits connected through long-range couplers, which were first tested in IBM’s experimental processors last year.
The tech giant said its community tracker currently supports three types of quantum advantage experiments: observable estimation, variational problems, and problems with classical verification.
To help developers manage the increasing complexity of quantum systems, IBM also announced updates to Qiskit, its open-source quantum software stack. The software now provides more precise control of circuits and improved performance through the use of high-performance classical computing for error correction.
The US-based tech company said enhancements to Qiskit’s dynamic circuit capabilities have improved accuracy by 24% on processors with more than 100 qubits. A new execution model and a C-API (application programming interface) will allow developers to integrate high-performance computing tools for faster error mitigation and reduced computational costs.
The company is also introducing a C++ interface to Qiskit to allow quantum programming directly within high-performance computing environments. By 2027, IBM plans to expand Qiskit with computational libraries for applications such as machine learning and scientific simulations.
On its roadmap toward fault-tolerant quantum computing, it announced IBM Quantum Loon, an experimental processor that showcases the components required for reliable, large-scale quantum systems. IBM said Loon will validate a new architecture for practical quantum error correction, including high-quality routing layers that connect distant qubits on the same chip.
The company also achieved real-time error decoding using classical computing hardware in less than 480 nanoseconds through qLDPC codes, a step completed a year ahead of schedule.
IBM said fabrication of its quantum processor wafers is now being done at the Albany NanoTech Complex in New York, using advanced semiconductor tools that have doubled the speed of development. The company said this has allowed it to increase the complexity of its chips tenfold and test multiple processor designs simultaneously.