Quantum computing is entering the enterprise through the cloud.
That matters because most organizations will not buy, install, or operate quantum hardware. Quantum computers simply aren’t data-center friendly. Instead, enterprises will access quantum processing units (QPUs) alongside CPUs, GPUs, and high-performance computing resources through familiar cloud environments.
Evaluating this market requires looking beyond physical qubit counts. While hardware architecture matters, so do software abstractions, classical-computing integration, error correction, geographic availability, and the partnerships connecting each platform to the wider quantum ecosystem.
The good news for most enterprise IT practitioners is that the public cloud environment you’re already dealing with is well aligned with advances in quantum computing.
The Hyperscalers
Amazon Web Services
Amazon Web Services (AWS)’s quantum play will be familiar to anyone already versed in its traditional infrastructure business: separate the application layer from the underlying hardware, reduce the need to commit capital to a single technology, and provide developers with a common environment for experimentation.
Amazon provides access to quantum computing through its Amazon Braket service, its multi-hardware development platform for QC:
- AWS provides access to multiple hardware modalities, allowing customers to explore different approaches without committing to a single vendor.
- Braket Hybrid Jobs connects quantum processors to classical Amazon EC2 resources, supporting the iterative classical–quantum workflows used by many current algorithms.
AWS Quantum Momentum
- In June 2026, AWS and QuEra Computing Inc. expanded their strategic collaboration and announced plans to make Libra available via Amazon Braket in 2028. Libra is QuEra’s megaqubit-class, fault-tolerant system targeting more than 256 error-corrected logical qubits and a logical error rate of 10⁻⁶.
- AWS continues to pursue proprietary, error-corrected superconducting hardware through its AWS Center for Quantum Computing while offering third-party systems via Braket. This gives AWS parallel internal-hardware and marketplace strategies (although AWS has not announced a public delivery date for its own fault-tolerant computer).
Microsoft Azure Quantum
Microsoft has one of the market’s most layered strategies. It operates a multi-vendor hardware marketplace, offers specialized software via Azure Quantum Elements, and maintains an internal quantum hardware program separate from the systems it hosts.
Microsoft’s early for Azure Quantum is on enterprise applications across chemistry, pharmaceuticals, and materials science:
- Azure Quantum Elements combines AI, classical high-performance computing, and quantum services. This enables Microsoft to deliver value from computational chemistry workflows today without making adoption dependent on fault-tolerant quantum hardware.
- Azure hosts hardware from partners such as Quantinuum, Atom Computing, IonQ, Rigetti Computing, and Pasqal.
- In parallel, Microsoft is developing topological qubit hardware, most visibly through its Majorana 1 processor. This gives the company a direct hardware stake and a marketplace role.
Microsoft Quantum Momentum
- Its ongoing collaboration with Quantinuum and, separately, with Atom Computing has repeatedly demonstrated the ability to extract stable, low-error logical qubits from physical hardware, most recently in the Atom Computing/Microsoft/QuNorth “Magne” system, which targets 50 logical qubits from more than 1,200 physical qubits by late 2026.
- In June 2026, Microsoft announced Majorana 2, its second-generation topological quantum chip. Microsoft reported that its qubits are 1,000 times more reliable than those in Majorana 1, with a mean qubit lifetime of 20 seconds, and says the advance supports a revised target of delivering a scalable quantum computer by 2029.
- Also in June, Microsoft highlighted peer-reviewed results from its collaboration with Quantinuum showing logical error-rate improvements ranging from 11× to 800× over corresponding physical-circuit baselines. The work included circuits spanning up to 12 logical qubits. Microsoft also released deq, an open-source error-correction package for its Quantum Development Kit.
- Microsoft and Atom Computing continue to develop Magne, a planned neutral-atom system targeting 50 logical qubits. QuNorth has reserved the system for deployment in Denmark.
IBM Quantum Platform
IBM takes a markedly different approach from traditional public cloud providers. Rather than operating a broad, multi-vendor marketplace, it centers its platform on IBM hardware and the Qiskit software ecosystem.
IBM treats quantum computing as a core layer of advanced computing infrastructure, anchored by its Heron and Nighthawk processors and the open-source Qiskit stack:
- As IBM has moved away from hosting general-purpose cloud notebooks, managed development environments from providers such as qBraid and Strangeworks have become key entry points for developers.
- IBM’s own cloud platform is increasingly focused on Qiskit Runtime, a high-throughput environment for running quantum workloads.
IBM Quantum Momentum
- In June 2026, IBM committed more than $10 billion over five years to quantum-computing R&D, capital investment, manufacturing expansion, ecosystem partnerships, and potential acquisitions. The investment is intended to support IBM’s existing roadmap and to accelerate progress beyond it.
- IBM’s 2026 roadmap calls for demonstrating initial examples of quantum advantage through integrated quantum and HPC workflows. For Nighthawk, IBM targets circuits with 7,500 gates and configurations with up to three connected 120-qubit modules in 2026. IBM also plans to prototype a real-time error-correction decoder this year.
- IBM continues to target a large-scale fault-tolerant system in 2029. This provides a useful comparison with AWS’s and QuEra’s 2028 Libra targets, even though the companies use different architectures and performance metrics.
Google Quantum AI
Google combines a research-led quantum program with an AI-first cloud strategy and the expansion of access to third-party hardware. Its quantum services remain more targeted than a general-purpose public utility, but its portfolio now spans both partner systems and internal hardware research.
Google’s strategy emphasizes quantum machine learning and the integration of quantum workflows into its AI and classical-computing infrastructure.
- Google connects its internally developed Willow superconducting processor with tools and environments including Cirq, TensorFlow Quantum, Vertex AI, and Cloud TPUs.
- Google Cloud also offers access to Pasqal’s neutral-atom systems through its marketplace, giving customers a third-party option beyond Google’s internal hardware.
Google Quantum Momentum
- In March 2026, Google Quantum AI expanded its hardware research beyond superconducting processors by launching a neutral-atom program led by JILA Fellow Adam Kaufman. Google says the two modalities offer complementary scaling characteristics: superconducting systems support faster, deeper circuits, whereas neutral-atom systems can provide large, reconfigurable qubit arrays.
- Google continues to develop Willow as a research platform for error correction and verifiable quantum computation. Its current access model is targeted and research-oriented, rather than publicly available as a quantum cloud service.
- Google also supports ecosystem development through Cirq, TensorFlow Quantum, research-access programs, and the XPRIZE Quantum Applications competition.
The Partnerships Shaping the Market
Cloud quantum computing depends on technical integration among cloud providers, hardware developers, software platforms, and enterprise service firms. Let’s look at a few of the most impactful.
AMD & IBM
IBM and AMD are collaborating on quantum-centric supercomputing architectures that combine IBM quantum processors with AMD CPUs, GPUs, and high-performance computing technology.
Announced in 2025 and expanded in 2026, the partnership is an architectural and research initiative rather than a generally available integrated product, but it highlights AMD’s growing role in the classical infrastructure required to make enterprise quantum workflows practical.
AWS and QuEra
This partnership anchors AWS’s neutral-atom strategy. The expanded agreement with Libra gives AWS a specific fault-tolerance target in its brokered hardware marketplace.
Microsoft Azure and Quantinuum
Microsoft combines Quantinuum’s trapped-ion hardware with its own logical-qubit and error-correction software. The partnership has repeatedly demonstrated low-error logical operations and complements rather than replaces Microsoft’s internal Majorana hardware program.
IBM, qBraid, and Strangeworks
As IBM has focused its native platform on compute execution through Qiskit Runtime, qBraid and Strangeworks have taken on a larger role in providing managed development and notebook environments. This division allows IBM to focus on its hardware and runtime infrastructure while partners reduce friction for developers.
NVIDIA CUDA-Q and the Multi-Cloud Ecosystem
Nvidia‘s CUDA-Q provides a bridge among GPUs, CPUs, and multiple types of quantum hardware. Its integrations with Amazon Braket, Azure Quantum, and infrastructure providers such as HPE make it an important framework for developing hybrid GPU–QPU workflows.
Enterprise Integrators
Accenture, Deloitte, Capgemini, IBM Consulting, and Microsoft Consulting Services help enterprises translate quantum experiments into governed programs. Their role includes selecting use cases, integrating quantum work into existing cloud environments, managing security and data requirements, and distinguishing technically credible opportunities from speculative ones.
Impact on Enterprise Architecture
The market has moved beyond the binary question, “When will a quantum computer outperform a classical one?”
The more useful model is hybrid computing. An enterprise workflow may ingest and transform data on CPUs, use GPUs for machine learning or dense numerical processing, and delegate a specialized optimization or molecular-simulation routine to a cloud-hosted QPU.
Classical and quantum systems will operate as components of the same workflow, not as competing, all-or-nothing alternatives.
For enterprise architects, this changes the planning horizon. The immediate task is to develop the skills, interfaces, governance, and workload selection methods needed to incorporate quantum resources when they become advantageous.
Analyst Recommendations for Enterprise Buyers
| Hyperscaler | Model | Hardware Partners | Software Partners | Native Software Stack |
| AWS (Amazon Braket) | Brokered Marketplace | QuEra (neutral atom), IonQ (trapped ion), Rigetti (superconducting), Oxford Quantum Circuits (superconducting) | NVIDIA CUDA-Q | Braket SDK / OpenQASM |
| Microsoft Azure Quantum | Hybrid Marketplace / Advanced Software Layer | Quantinuum (trapped ion), Atom Computing (neutral atom), IonQ (trapped ion), Rigetti (superconducting), Pasqal (neutral atom) | NVIDIA CUDA-Q | Q# / Azure Quantum SDK |
| IBM Quantum Platform | Vertically Integrated | None — proprietary Heron / Nighthawk fleet only | qBraid, Strangeworks (managed notebook environments) | Qiskit Runtime |
| Google Cloud Quantum | Vertically Integrated (AI Focus) | Pasqal (neutral atom); internal Google Quantum AI (Sycamore, Willow) | Native tooling | Cirq / TensorFlow Quantum |
There is a growing number of cloud-based options to ease an enterprise’s expansion into quantum computing, but how should enterprises approach quantum overall?
Select platforms by workload, not headline qubit counts
For materials science, chemistry, biopharmaceutical research, and energy-storage applications, the software and domain-modeling layer may matter more than the underlying processor.
Look for platforms that can deliver near-term value by combining AI, high-performance computing, and quantum-oriented services while shielding researchers from certain low-level implementation details.
Evaluation criteria should include:
- Relevance to the target workload
- Quality and maturity of development tools
- Integration with existing cloud and HPC environments
- Access to multiple hardware modalities
- Runtime performance, queue times, and pricing
- Error-mitigation and error-correction capabilities
- Data residency, security, and compliance controls
- Portability of code, data, and intellectual property
Build a portfolio of experiments
A credible enterprise quantum program should not begin with a mandate to move applications into production. It should begin with a small portfolio of testable use cases; each paired with a strong classical benchmark.
Teams should define in advance what constitutes meaningful progress: improved solution quality, reduced runtime, lower energy consumption, access to a previously intractable problem, or new scientific insight. Without such benchmarks, a pilot can demonstrate technical activity without establishing business value.
Prepare for security and sovereignty requirements
Quantum planning is also a security and governance issue. U.S. National Security Memorandum 10 and related federal guidance have established timelines for migrating vulnerable cryptographic systems toward post-quantum cryptography.
European data-residency and sovereignty requirements add another layer of complexity for regulated workloads.
Enterprises in banking, defense, healthcare, and critical infrastructure should evaluate:
- Where quantum workloads and associated data will be processed
- Which parties can access algorithms, results, and intellectual property
- Whether regional or sovereign deployment options are available
- How the provider supports cryptographic agility
- How quantum services fit into existing third-party risk and cloud-governance programs
European options include Scaleway deployments that incorporate AQT trapped-ion and Pasqal neutral-atom hardware, while IBM operates quantum infrastructure through several regional partnerships and installations.
The Next Enterprise Computing Skill Is Orchestration
Fault-tolerant quantum computing has not yet reached enterprise scale, and vendor roadmaps should be treated as targets rather than guarantees. However, the cloud platforms, software layers, and development ecosystems forming around quantum computing are already usable for research, workforce development, and carefully scoped experimentation.
That makes the next step more practical than it may seem. Enterprise teams don’t need to choose the eventual winner in quantum hardware; rather, they should identify valuable workloads, establish classical benchmarks, gain experience across multiple architectures, and implement appropriate security and governance controls.
The enterprises best positioned to benefit from quantum computing will be those that learn to orchestrate CPUs, GPUs, and QPUs as components of a unified computing environment.
This allows them to recognize when a quantum resource has earned a place in a production workflow, perhaps the most critical skill needed for enterprise adoption of the technology.
