IBM has launched a new service to help companies deploy quantum computers and train staff to use the emerging technology.

Called the Quantum Accelerator, the program will involve IBM experts working with companies to figure out the most valuable near-term use cases for quantum computers in their business and then help them research and deliver proof-of-concept projects around those use cases. It will also involve the development of what Katie Pizzolato, IBM’s head of quantum strategy, said are bespoke curriculums to train a customer’s staff on how to use quantum computers and quantum algorithms.

The company said it would tailor its educational offering to the needs of different groups within each company. Some high-level executives might need a basic grounding in what quantum computers are and how quantum algorithms differ from conventional ones and where these might help with business needs. Meanwhile, software engineers might require far more technical training in how to program quantum computers or write their own quantum algorithms.

IBM is among dozens of companies racing to commercialize quantum computing. Big Blue built one of the first working quantum computers, but it has now been joined by others that include rival tech giant Google, electronics powerhouse Honeywell, and many startups, including Rigetti Computing and IonQ. There are many more companies also competing with IBM to develop quantum algorithms and software for quantum computers.

In theory, quantum computers are exponentially more powerful than conventional ones. Conventional computers, which computer scientists call classical computers, can only process information in a binary format, called a bit, which is either a 0 or 1. Quantum computers harness the properties of quantum mechanics to create processing units called qubits, which can represent both 0 and 1 simultaneously. In addition, the state of a bit in a traditional computer is designed to be independent of all the other bits in the circuit. But in a quantum computer, the qubits are designed to affect one another, so they can all process a problem in parallel, rather than in a linear fashion.

The problem with today’s quantum computers is two-fold: they have relatively few qubits—the largest quantum processors IBM currently produces have about 65—and those qubits can remain in a quantum state for relatively short periods of time. And when those qubits fall out of a quantum state, they produce errors. Those errors must then be corrected with yet more physical qubits or various kinds of software. At the moment, those errors can negate the theoretical advantage of using a quantum computer.
 
IBM has announced plans to have a quantum computer with more than 1,000 qubits available by 2023. The company has said a machine with that kind of power should be useful for many real-world business problems, such as chemical simulations or better calibrating the tradeoffs between potential risk and reward in some financial portfolios. It plans to have a machine with 1 million qubits available by 2030.

But today’s announcement of The Accelerator seems to be an acknowledgement that there won’t be a ready market for its 1,000 qubit machines in 2023 unless IBM actively tries to seed it now by getting more customers ready to deploy the technology. (Pizzolato said there was no problem with adoption of IBM’s current quantum offerings but that the company wanted to offer businesses training and help thinking through how to use quantum computers “in parallel with the roadmap” on quantum hardware.)

The Accelerator will also make it more likely that customers will chose IBM’s quantum hardware and software over competitors’ since they will be more familiar with it from the tutorials and research projects.

IBM may be concerned because its approach to quantum computing is based around qubits formed by superconducting materials. But other companies, including a number of startups, are using technology based on ions trapped with powerful lasers. And those qubits have so far proven to be less error-prone than superconducting qubits. So until larger systems, with more error-correction, are built, these other kinds of qubits might be more attractive for business customers.

The Accelerator is also another way for IBM to monetize its investment in quantum know-how in the near-term. Pizzolato declined to discuss how much the quantum accelerator service will cost, but she said customers would have to pay for access to the program.