Quantum computing relies on specialized hardware to create and manipulate qubits. The development of this hardware is a major area of research and engineering, with various approaches being explored by different companies and institutions.
Types of Qubits and Processors
Different physical systems are used to realize qubits, each with its own advantages and challenges in terms of scalability, coherence, and connectivity.
- Superconducting Qubits: Many companies, including IBM and Google, use superconducting qubits. These require extremely low temperatures, colder than outer space, typically achieved using dilution refrigerators. IBM has developed processors like the 433-qubit IBM Quantum Osprey and the 133-qubit IBM Quantum Heron. Google has also developed state-of-the-art chips like Willow, described as a big step towards large-scale, error-corrected quantum computing.
- Trapped Ions: This approach uses charged atoms held in place by electromagnetic fields. Companies like Quantinuum use ion-trap hardware.
- Neutral Atoms: Companies like Pasqal specialize in neutral atom technology. Researchers have achieved high two-qubit gate fidelity on neutral atom qubits.
- Topological Qubits: Microsoft is pursuing topological qubits, aiming for a new kind of qubit based on a special material called a topoconductor, which can create a topological state of matter. Their Majorana 1 processor uses this approach and is designed with error resistance at the hardware level. This architecture aims for a clear path to fitting a million qubits on a single chip.
- Photonic Qubits: Xanadu and PsiQuantum focus on photonic quantum computing.
- Cat Qubits: Pursued by Alice & Bob, these qubits aim for protection against bit flips.
Architecture and Scaling
A major challenge in quantum hardware is scaling the number of qubits while maintaining their coherence and controlling their interactions. Microsoft envisions a path to a million qubits. Google is targeting a million qubits by the end of the decade. IBM aims to build a 100,000-qubit machine within 10 years and is developing modular and extensible systems leveraging high-performance classical computers (HPCs) for quantum-centric supercomputing. Their roadmap focuses on increasing both the number of qubits and “gate operations,” a measure of workload scale. IBM is also developing a 4K cryo-CMOS qubit controller to control qubits from inside the refrigerator.
The path to large-scale, fault-tolerant quantum computing requires overcoming engineering challenges such as cooling power, individual qubit control at scale, speed of computation, connectivity, and manufacturability. The goal is to transition from the current noisy intermediate-scale quantum (NISQ) era toward resilient systems where logical qubits are protected from noise.
Quantum chips do not work in isolation. They exist within an ecosystem including control logic, cryogenic systems (for some modalities), and a software stack.
