The Difference Between Noisy and Useful
The quantum computing industry has spent the past decade operating primarily in what researchers call the noisy intermediate-scale quantum era. NISQ devices are quantum processors containing between fifty and a few thousand physical qubits that are too error-prone to run the deep quantum circuits that the most commercially valuable quantum algorithms require. The noise that characterises NISQ devices comes from the fragility of the quantum states that qubits must maintain during computation. Any interaction between a qubit and its environment, including thermal vibrations, electromagnetic interference, and even cosmic rays, can cause the qubit to lose its quantum state in a process called decoherence. At current physical qubit error rates, the errors accumulate faster than meaningful computation can proceed for the deep circuits that commercially significant quantum algorithms require. NISQ devices have demonstrated genuine quantum advantage over classical computers on specific artificial benchmarks, but the practical commercial applications that require the sustained, error-free quantum computation that fault-tolerant quantum processors provide have not been accessible on NISQ hardware.
Fault-tolerant quantum computing uses quantum error correction codes to create logical qubits whose error rate is substantially lower than the physical qubit error rate through the redundant encoding of each logical qubit across many physical qubits. A fault-tolerant quantum processor can run arbitrarily deep quantum circuits with controlled error rates by continuously detecting and correcting errors in the physical qubit layer without disrupting the logical computation. The resource overhead of quantum error correction is enormous. A single logical qubit with an error rate suitable for commercially relevant computation requires thousands of physical qubits with current error correction codes at current physical qubit fidelities. Building a fault-tolerant quantum processor with enough logical qubits to run commercially significant algorithms therefore requires either dramatically improving physical qubit fidelity, developing more efficient error correction codes, or building processors with very large numbers of physical qubits. All three approaches are being pursued simultaneously by the leading quantum hardware companies.
IBM Heron and the Processor Architecture Race
IBM's quantum roadmap has been the most publicly detailed of the major quantum hardware programmes and provides the clearest public reference for the pace of progress in superconducting qubit quantum processor development. IBM's Heron processor, released in late 2023, represents a significant improvement in two-qubit gate error rates over the earlier Eagle and Osprey processors, achieving the physical qubit fidelities that make quantum error correction increasingly practical. IBM's stated roadmap targets fault-tolerant quantum computing with error-corrected logical qubits within this decade, and the engineering progress visible in the Heron architecture's reduced coupling errors and improved gate speeds provides technical credibility for the roadmap that its predecessors were less able to demonstrate. The commercial significance of IBM's quantum programme extends beyond its own quantum hardware to the IBM Quantum Network of industry partners whose access to IBM quantum systems and the Qiskit software stack creates the developer ecosystem that commercial quantum applications require.
Google's Willow quantum chip, announced in late 2024, demonstrated below-threshold quantum error correction, meaning that the logical qubit error rate decreased as more physical qubits were added for error correction rather than increasing as had previously been observed. This threshold crossing is the fundamental physical demonstration that quantum error correction can achieve the error rate reduction that fault-tolerant quantum computing requires, and its demonstration on a superconducting qubit processor with 105 qubits represents the most significant experimental quantum computing milestone since the earlier quantum supremacy demonstrations. The commercial implications of Willow's below-threshold error correction are not immediate but are directionally significant: the physical evidence now supports the engineering trajectory that leads to commercially useful fault-tolerant quantum computation.
Microsoft's Topological Qubit Approach
Microsoft's approach to fault-tolerant quantum computing differs fundamentally from the superconducting qubit architectures of IBM and Google. Topological qubits, which encode quantum information in the global topological properties of a physical system rather than in the state of individual particles, are theoretically more robust to local perturbations that cause decoherence in conventional qubit types. Microsoft's Majorana 1 chip, announced in early 2025, demonstrated the creation of topological qubits using a new class of material called a topoconductor, whose properties allow the formation of Majorana zero modes that are the physical basis for topological qubit encoding. If the topological approach achieves its theoretical error rate advantage, it would allow fault-tolerant quantum computation with far fewer physical qubits per logical qubit than error correction codes applied to superconducting or trapped ion qubits require. The commercial consequence of a successful topological qubit approach would be a more resource-efficient path to fault-tolerant quantum computation that reaches commercially useful logical qubit counts faster than the alternative approaches whose error correction overhead is substantially larger.
Top 10 Companies in Quantum Computing Globally
- IBM Quantum: Most commercially deployed quantum computing programme with the largest quantum developer ecosystem through Qiskit and IBM Quantum Network; its Heron processor architecture and its transparent public roadmap toward fault-tolerant quantum computing make it the reference programme against which commercial quantum timelines are assessed.
- Google Quantum AI: Willow chip's below-threshold quantum error correction demonstration is the most significant experimental milestone in fault-tolerant quantum computing; its focus on superconducting qubit processor development and its Cirq software platform are positioned for the transition from NISQ demonstration to fault-tolerant application development.
- Microsoft Azure Quantum: Topological qubit approach through Majorana 1 chip represents the highest-risk highest-reward quantum hardware bet; its Azure Quantum cloud platform and its Q# programming language are the commercial infrastructure whose value is independent of whether the topological hardware approach achieves its theoretical advantages on the timelines Microsoft's roadmap implies.
- IonQ: Publicly listed trapped ion quantum computing company whose qubit fidelity advantage over superconducting systems at current qubit counts is its primary commercial differentiator; its algorithmic qubit metric and its cloud access through AWS, Azure, and Google Cloud create the multi-cloud distribution that maximises its addressable customer base.
- Quantinuum: H-series trapped ion quantum processors with the highest published two-qubit gate fidelities of any commercial quantum system; its System Model H2 with 56 qubits achieves error rates that make quantum error correction genuinely practical and its Quantum Charge Coupled Device architecture is the trapped ion approach closest to fault-tolerant operation.
- Amazon Braket / AWS Center for Quantum Computing: Cloud quantum computing service providing access to multiple quantum hardware technologies alongside Amazon's own quantum hardware development programme; its commercial significance lies more in the cloud distribution infrastructure than the hardware development, creating the enterprise customer onboarding path that quantum hardware companies depend on.
- QuEra Computing: Neutral atom quantum computing company whose Aquila system demonstrated programmable quantum simulation with 256 qubits; its reconfigurable atom array approach and its Harvard University research partnership create a neutral atom hardware development programme that is advancing alongside the superconducting and trapped ion mainstream.
- PsiQuantum: Photonic quantum computing company developing fault-tolerant quantum computers using silicon photonics manufacturing; its fabrication partnership with GlobalFoundries and its focus on photon-based logical qubits represent the approach most dependent on semiconductor manufacturing scale to achieve the million-physical-qubit counts its fault-tolerance architecture requires.
- Xanadu: Canadian photonic quantum computing company whose Borealis system demonstrated quantum computational advantage on photonic hardware and whose PennyLane software framework is the most widely used quantum machine learning development platform; its open-source software strategy creates the developer community that commercial quantum software companies need before hardware matures.
- Rigetti Computing: Publicly listed superconducting quantum computing company with its own fab for quantum chip manufacturing; its Ankaa processor and its Quantum Cloud Services platform serve the research and early commercial quantum application market while the fault-tolerant era matures.