With atoms and lasers, IonQ makes a quantum computing leap

IonQ expects to deliver by year's end its first cloud-based services from what it claims will be the industry's fastest quantum computer.

The startup, based in College Park, Md., recently released the results of two benchmark tests published on Cornell University's Quantum Physics site that showed its upcoming system is capable of solving more complex problems with a higher degree of accuracy than any other results published by competitive systems.

What is particularly interesting about the IonQ system is the distinctly different approach the company takes in developing the system compared with that of competitors such as IBM, Microsoft, Rigetti and Google.

The qubits used in the IonQ system are individual atoms of the element ytterbium that store the information. The information can be processed and retrieved from the atoms using sophisticated lasers, a method  company executives refer to as ion trapping.

Companies like IBM and Google use silicon chips to create their qubits, which require expensive control systems to keep temperatures close to absolute zero. The IonQ system's technology, however, allows the machine to operate in room temperatures. And while silicon has the advantage of being a mature technology, vendors are still struggling to make their systems more stable and to reduce the number of errors caused by noise.

IonQ officials said its system also has issues with noise, but those issues largely have to do with the mechanical parts of the machine's control systems. Company executives believe those issues can be resolved through a straightforward trial-and-error process by isolating them down one by one.

"In all the places in the system where we have noise and errors, they are known to us," said Stewart Allen, IonQ's founder and chief product officer. "There is nothing involving the quantum computing technology we are trying to improve upon. It's all in the control systems."

Because the system doesn't require a temperature-controlled environment, IT shops can bring it on premises. But Allen said he doesn't expect that to happen for some time and initially will deliver quantum capabilities to users as cloud-based services.

There are two major components of the system that are each controlled by the software: the laser beams and the trap or chip, according to Allen. The chip provides an electromagnetic bottle that holds the charged ions. The ions communicate through the coulomb force that repels each other because of their charge. This allows the chip to trap them and holds them suspended in a vacuum. It is this isolation that allows the ions to retain quantum states longer than competing technologies, Stewart explained.

The IonQ system can store up to 160 qubits and carry out operations on 79 qubits, Allen said. He added the system has a "gate fidelity," which measures the accuracy of logical operations, of 98% for one- and two-qubit operations in a 13-qubit configuration. This level of performance allows it to handle longer calculations.

IonQ was formed in 2015, but only began emerging from stealth mode over the past six months. But the company has made significant progress over the past couple of years in terms of putting all the pieces of the system together.

Going up against companies like IBM and Google ... it's hard to convince people you are doing something better. So, we thought it wiser to get the performance benchmarks on the record and then show the market, 'We have done better, and here it is.'

"Two years ago, we had just received funding, had no offices, employees or computers," Allen said. "But, in that time from a standing start, we got the facilities, built the computers and can operate them at a very high level. I think that sort of progress validates the approach we are taking," he said.

Explaining why the company has moved forward slowly, Allen said he didn't want to debut the technology until it could fully demonstrate its core capabilities, given the level of competition it was up against. Overhyping an unproven product would handicap its chances right out of the starting blocks.

"Going up against companies like Google and IBM that have been doing this a long time, it's hard to convince people you are doing something better," Allen said. "So, we decided it was wiser to get the performance benchmarks on the record and then come out and show the market, 'We have done better, and here it is,'" he said.

IonQ has built three systems and expects to build more by the end of this year. Initially, the company will provide quantum-based services to researchers using applications for medicine, chemistry, energy, logistics and those working in the financial world, according to Allen. In short, it's for applications where conventional computers can't offer adequate speed and performance for data-intensive tasks, such as simulating models for quantum physics or optimizing complex cost functions.

The product has attracted "a lot of tire kickers," Allen said, but the company will continue to move forward cautiously. Over the next few months, the company will invite larger companies in the private sector, as well as selected government agencies, to put the system through its paces and hopefully collect some constructive feedback.

"We are approaching the point where we can do some interesting things with certain types of molecular simulations and perhaps simulations around problems involving financing and logistics," Allen said.