Since late 2019, Amazon Braket has been in private preview. However, Amazon AWS recently announced its general availability. Customers can purchase quantum computing services directly from the Amazon Braket platform.
Amazon named the service after a standard quantum notation called Bra–ket that was created in the late 1930s by Paul Dirac, a famous theoretical physicist.
Braket is a fully managed AWS service that provides a mostly technology-agnostic environment. It offers customers the capability to design, develop, test, and run their quantum algorithms. Braket also offers simulated quantum computers that use Amazon EC2 computing resources and can simulate up to 34 qubits for testing and troubleshooting algorithms.
For now, the service is currently available in the following AWS regions: US East (N. Virginia), US West (N. California), and US West (Oregon).
Braket offers a choice of three different quantum hardware technologies.
Each piece of quantum hardware can be accessed via Braket but resides on each respective provider’s premises. Once a customer is satisfied that their quantum algorithms are complete, they can run them on their choice of three different types of quantum hardware.
The Amazon Braket quantum hardware providers are:
- Rigetti Computing provides a gate-based, quantum processor using 30 superconducting qubits that operate at cryogenic temperatures. Superconducting qubits are a mature technology with easy coupling, electronic control, and fast gates. It is the most commonly used quantum technology.
- IonQ provides a gate-based trapped ion quantum computer with 11 qubits. Individual ions, held in a vacuum via an electromagnetic trap generated by surrounding electrodes, use laser pulses to perform gate operations. It has relatively long coherence times and excellent connectivity.
- D-Wave runs a quantum annealing processor with 2048 qubits. Quantum annealing uses a physical process to find a low energy configuration that encodes the solution of an optimization problem. Note: You cannot make a qubit-to-qubit comparison between quantum annealers and gate-based machines. They are different technologies.
Braket’s pricing structure has two components – a per task fee and a per shot fee. There is also an hourly rate for using the Amazon Braket managed quantum simulator.
Amazon Braket defines a task as repeated shots (runs) based on the same circuit design or annealing problem. There is a level per-task pricing for D-Wave, Rigetti, and IonQ.
Shots have variable pricing. Think of a shot as each time an algorithm runs. D-Wave, Rigetti, and IonQ all have different per shot rates.
Keeping it in perspective
Amazon Braket isn’t the only quantum cloud service.
IBM has a massive fleet of 22 superconducting quantum computers. In 2016, IBM was the first company to put a quantum computer online. In addition to offering a premium service, most of IBM’s quantum systems are publicly available, along with complete tutorials and a large library of useful algorithms. According to IBM, billions of quantum circuits are run each month on its network.
Microsoft Azure Quantum offers a similar cloud service to Braket. However, its platform is still in preview by a select group of customers. It is not available for general use. The Azure platform offers diverse trapped ion technologies from Honeywell Quantum Solutions and IonQ, plus superconducting hardware from Quantum Circuits, Inc. (QCI). Honeywell recently announced it has the most powerful quantum computer in the world, with a quantum volume of 64.
Although the quality of qubits is an important factor, the number of qubits by itself doesn’t tell the entire performance story. There is no current standard performance benchmark for quantum computers. In the interim, quantum volume is an appropriate measurement for gate-based quantum computers. It is hardware-agnostic and takes into account the number of qubits, connectivity, and gate and measurement errors. Additionally, and equally important, its numerical simplicity makes quantum computing more understandable to the general public, corporate management, and investors. Unfortunately, although it would be interesting, very few companies are publishing their quantum volume numbers.
Keep in mind that we are still in an experimental era of quantum computing. For the most part, a classical computer can do anything a quantum computer can do.
Still, quantum cloud computing provides expanded global access to the technology. It allows customers, students, and researchers to begin learning and planning for the future.
There will be a time, perhaps in 10 to 20 years, maybe sooner with some technologies we have seen on roadmaps, when quantum applications will be an integral part of our daily lives. Then, quantum computers will have the capability to solve major scientific, financial, and medical problems of tomorrow.
Note: This analysis contains content from Moor Insights & Strategy President & Principal analyst Patrick Moorhead.
Disclosure: Moor Insights & Strategy, like all research and analyst firms, provides or has provided paid research, analysis, advising, or consulting to many high-tech companies in the industry, including 8×8, Advanced Micro Devices, Amazon, Applied Micro, ARM, Aruba Networks, AT&T, AWS, A-10 Strategies, Bitfusion, Blaize, Calix, Cisco Systems, Clear Software, Cloudera, Clumio, Cognitive Systems, CompuCom, Dell, Dell EMC, Dell Technologies, Diablo Technologies, Digital Optics, Dreamchain, Echelon, Ericsson, Extreme Networks, Flex, Foxconn, Frame, Fujitsu, Gen Z Consortium, Glue Networks, GlobalFoundries, Google (Nest-Revolve), Google Cloud, HP Inc., Hewlett Packard Enterprise, Honeywell, Huawei Technologies, IBM, Ion VR, Inseego, Intel, Interdigital, Jabil Circuit, Konica Minolta, Lattice Semiconductor, Lenovo, Linux Foundation, MapBox, Mavenir, Marseille Inc, Mayfair Equity, Meraki (Cisco), Mesophere, Microsoft, Mojo Networks, National Instruments, NetApp, Nightwatch, NOKIA (Alcatel-Lucent), Nortek, Novumind, NVIDIA, ON Semiconductor, ONUG, OpenStack Foundation, Oracle, Poly, Panasas, Peraso, Pexip, Pixelworks, Plume Design, Portworx, Pure Storage, Qualcomm, Rackspace, Rambus, Rayvolt E-Bikes, Red Hat, Residio, Samsung Electronics, SAP, SAS, Scale Computing, Schneider Electric, Silver Peak, SONY, Springpath, Spirent, Splunk, Sprint, Stratus Technologies, Symantec, Synaptics, Syniverse, Synopsys, Tanium, TE Connectivity, TensTorrent, Tobii Technology, Twitter, Unity Technologies, UiPath, Verizon Communications, Vidyo, VMware, Wave Computing, Wellsmith, Xilinx, Zebra, Zededa, and Zoho which may be cited in this article.