You don’t need a quantum computer to start learning quantum computing. The major cloud platforms give pay-as-you-go access to real quantum hardware and high-performance simulators. Here is how the main options compare and how to plan a first experiment.
The main platforms
- Amazon Braket — AWS’s managed quantum service, offering access to hardware from several third-party providers using different qubit technologies, plus managed simulators and hybrid jobs that run classical and quantum code together.
- Azure Quantum — Microsoft’s quantum platform, providing access to partner hardware, development tools and resource estimation for planning future fault-tolerant algorithms.
- IBM Quantum — access to IBM’s own superconducting processors through the cloud, closely tied to the widely used open-source Qiskit SDK.
Available hardware, regions and pricing change frequently, so check each provider’s current offering before planning experiments.
How they compare
| Amazon Braket | Azure Quantum | IBM Quantum | |
|---|---|---|---|
| Hardware | Multiple third-party providers | Multiple partner providers | IBM processors |
| Main SDKs | Braket SDK, also works with Qiskit and PennyLane | Q#, Qiskit, Cirq support | Qiskit |
| Strengths | AWS integration, hybrid jobs | Azure integration, resource estimation | Large community, learning resources |
| Best fit | AWS-centric teams | Microsoft-centric teams | Learning and research |
Simulators first, hardware second
Most development happens on simulators, which are cheaper, faster and noise-free (or can simulate noise). Use real hardware to understand the effect of noise and to validate small circuits. Keep hardware runs small and budgeted — costs are usually charged per task and per shot.
A first project in four steps
- Choose a small, real problem — for example a tiny routing or scheduling instance from your business.
- Set a classical baseline with a good classical solver so you can judge results honestly.
- Prototype on a simulator using an open SDK such as Qiskit or PennyLane, then run a few hardware tasks.
- Write up the findings — what worked, what didn’t, how results scale and what it cost.
Avoiding lock-in
Open-source SDKs like Qiskit, Cirq and PennyLane run across several platforms. Writing experiments in them keeps your code portable as hardware and providers evolve.
Our Quantum Use-Case Lab runs these experiments with your team on the platform that fits your cloud strategy.
Frequently asked questions
Can I access a real quantum computer in the cloud?
Yes. Amazon Braket, Azure Quantum and IBM Quantum provide cloud access to real quantum hardware as well as simulators, typically on a pay-as-you-go basis.
Which quantum SDK should I learn?
Qiskit is the most widely used and works with IBM Quantum and other platforms; PennyLane is popular for quantum machine learning; Cirq is Google’s open-source framework. Open SDKs keep your work portable.
How much does cloud quantum computing cost?
Simulators are inexpensive for small circuits; real hardware is usually charged per task and per shot. Small, well-planned experiments can be run on a modest budget — check current provider pricing.