Wondering how you can run a real quantum circuit without owning a machine? Quantum cloud computing lets you reach remote quantum processors over the internet, through an account and a few lines of code, often for free.
This guide is part of our wider coverage of quantum technology, and it explains how remote access to quantum machines works in practice.
| Item | Value |
|---|---|
| IBM Quantum Experience launch | 2016 |
| Rigetti programmable cloud access demonstration | Early 2017 |
| Hardware makers on Amazon Braket (as of June 2025) | IonQ, Rigetti, IQM and QuEra |
| Amazon Braket on-demand simulators | 2 (SV1 and DM1) |
| QC Ware Forge free trial | 30 days, including one minute of quantum computing time |
These figures come from Wikipedia’s cloud-based quantum computing page and the Amazon Braket documentation overview. Providers change hardware lists, plans and trials often, so check the source before you rely on them.
Why the cloud suits quantum hardware
A quantum processor is a shared, finicky instrument. According to a Berkeley Nucleonics guide, a superconducting machine needs a dilution refrigerator the size of a small car and a room of supporting electronics. Nobody is going to put one on your desk.
Quantum work is also bursty. You spend most of your time writing and simulating on a classical computer, then send a compact description of a circuit to the quantum machine, which runs it many times and returns measurement outcomes. The same guide notes the processor is busy for a tiny fraction of your working time, so sharing it across thousands of users is both economical and technically sensible.
A January 2026 arXiv study of cloud quantum computing puts the shift plainly: the field has moved from a few custom machines open only to their creators to commercial machines that can be accessed on the cloud by anyone.

How cloud access began
The first public machines arrived only a few years before the current platforms. As Wikipedia’s overview records, IBM launched the IBM Quantum Experience in 2016, one of the first publicly accessible quantum processors connected to the cloud. In early 2017, researchers at Rigetti Computing demonstrated programmable cloud access through their Forest platform and its pyQuil Python library.
The same page says that since the early-2020s the field has grown significantly, with multiple providers offering a variety of hardware types:
- superconducting qubits
- trapped ions
- neutral atoms
- photonic systems
Who offers access, and through what tools?
Access comes in two broad shapes. Aggregator platforms such as Amazon Braket, Azure Quantum and qBraid bring devices from several hardware developers into one interface, according to Wikipedia. The same page says that, as of June 2025, Amazon Braket provides access to machines built by IonQ, Rigetti, IQM and QuEra.
The second shape is the maker’s own portal. Wikipedia notes that a lot of hardware manufacturers provide their own platform to reach their dedicated hardware.
These platforms usually support open-source SDKs, which are the code libraries you write programs in. Wikipedia names Qiskit, Cirq and PennyLane as examples. Some tools sit on top of the providers: Classiq’s documentation says it integrates with multiple cloud providers so programs can run on several backends, though some hardware providers may require an account.
Smaller offers exist too. Wikipedia lists a QC Ware platform, Forge, that gives access to D-Wave hardware and offers a 30-day free trial including one minute of quantum computing time.
Simulators, SDKs and hybrid jobs
Most of your early work happens on a simulator, not a real processor. The Amazon Braket documentation describes a local simulator in its SDK plus two on-demand simulators: SV1, a general-purpose state vector simulator, and DM1, a density matrix simulator that supports noise modeling.
Braket’s pages also describe several other building blocks that are typical of cloud platforms:
- Technology-agnostic SDK: it is designed so you do not have to code against each machine’s own programming environment.
- Two ways to run: on demand, or through a device reservation with Braket Direct.
- Hybrid Jobs: you provide an algorithm script and choose a quantum processor or simulator, and the service can spin up classical compute, run the algorithm and release the resources afterwards.
This hybrid pattern matters because many algorithms loop between a classical computer and a quantum one. The cloud service handles the classical half so you do not have to keep it running yourself.
What to check before you run a job
The arXiv study describes a complex chain of tools for connecting, programming, simulating, estimating resources, submitting jobs and retrieving results. It adds that some steps depend on the hardware and may change as gate sets and optimizing compilers evolve. Its own data covers machine accessibility, job submission times and success rates, circuit performance and cost.
The authors also report that circuit transpiling methods significantly change the outcomes of quantum circuits. That is a good reason to test before you spend money on a real device. A sensible order looks like this:
- Pick an SDK and write a small circuit.
- Run it on a simulator, with a noise model if one is offered.
- Estimate cost first. Classiq’s documentation, for example, points users to a cost estimation page before running.
- Choose a device, submit the job and retrieve the measurement results.
- Compare the hardware output with the simulated one.
If you are curious what people actually try on these machines, our guide to quantum computing applications is the next stop. Remote access is also a separate topic from encryption risk, which our page on post-quantum cryptography covers.
The Berkeley Nucleonics guide gives a useful warning for as of 2026: treat specific numbers, plan names and processor generations as a snapshot, because the structure of the ecosystem is more stable than any single product page.
Frequently asked questions
Do I need to own or visit quantum hardware?
Show answer
No. The hardware stays in controlled laboratories, and you reach it over the internet through an account and code, as the Berkeley Nucleonics guide describes.
Can I try it for free?
Show answer
Often, yes. The same guide says you can run a real circuit on a real processor, often for free, though plans differ by provider, so read each provider’s current terms.
Which programming tools can I use?
Show answer
Platforms often support open-source SDKs such as Qiskit, Cirq and PennyLane, according to Wikipedia. Amazon Braket also offers its own SDK, built to be technology agnostic.
Why simulate before using real hardware?
Show answer
Simulators let you test an algorithm cheaply before sending it to a device. Amazon Braket’s documentation notes that its DM1 simulator supports noise modeling, which helps you anticipate imperfect hardware.



