The chip will provide the scientific community with 16 qubits for exploring potential applications of the technology. According to IBM, quantum computers could dramatically advance fields such as machine learning and scientific modelling by increasing the amount of processing power available to researchers. Such systems’ strength comes from qubits’ ability to hold more information than the bits in a conventional machine, which makes it possible to perform certain operations considerably faster.
Ars Technica estimates that IBM’s new 16-qubit processor is about 40 percent more powerful than its 5-qubit predecessor. The comparison is based on the difference in their “quantum volume,” a new unit of measurement that the company has developed to evaluate quantum systems’ performance. It was conceived in response to the shortcomings of older evaluation methods, which have proven fairly ineffective at comparing different types of implementations.
IBM aims to provide a more objective benchmark by accounting for a wide range of factors including the fidelity of the qubits in a system, their number and circuit connectivity. If the approach lives up to the promise, it could be a major boon for the scientific community. And IBM’s commercialization would should benefit as well.
The company is working to develop production-grade quantum computers for powering chemistry models, artificial intelligence projects and other cutting-edge research. IBM will need a reliable means of comparing the systems against the competition when pitching potential clients, a requirement that the new assessment method could fulfill.
The fact that the company is starting to sort out the nuances of its competitive strategy reflect just how fast quantum computing is maturing. At the current rate of development, the technology could be ready for prime time in the foreseeable future.
In conjunction with the launch of the 16-qubit system, IBM revealed the existence of an even more sophisticated 17-qubit prototype (pictured) that is expected to be “at least twice” as powerful. Dario Gil, the head of the company’s Science and Solutions group, told Motherboard that the plan is to sell time on the hardware similarly to how it offers cloud infrastructure. IBM expects to produce its first 50-qubit machines a few years later.