Antoinette Brugman
3 July 2020

Though recently developed prototypes still consist of 50 qubits maximum, Delft startup Qblox is paving the way for building quantum computers with up to a 1,000 qubits and beyond.

Quantum computing is a promising technology that is on its way to create a new generation of powerful supercomputers. In the near future, these look to take on complex problems that no classical computer could ever solve. To achieve this, a quantum computer uses quantum mechanical phenomena and calculation units called qubits. Instead of ‘normal’ computer bits that can be either 1 or 0, a qubit can be 1 and 0 at the same time – so-called superposition. Moreover, qubits can be entangled. This means that two or more of them are correlated and an operation on one can instantaneously affect the other(s), regardless of the distance between them.

At the moment, quantum computers require rather large setups. The qubits are positioned on a chip that’s placed in a cryostat that cools down to a temperature of about 10 mK – near absolute zero. Such a low temperature is essential to avoid noise in the system caused by heat, for this would result in the qubits losing their information. The cooling part of the installation already takes up a lot of physical space.

Qblox team at work
In 2018, Qblox cofounders Jules van Oven (left) and Niels Bultink (right) set out to create a new system to control the qubits in a quantum computer. Credit: Qblox

However, the system that controls and reads out the qubits – an important part of the quantum computer – currently requires quite some room as well. As research in the domain of quantum computers is still in an experimental phase, this part of the installation is usually built using existing electronics. These systems are bulky, not specifically equipped for their task in a quantum computer setup and have scalability issues.

Saving space

Qblox saw a chance there. The spinoff of the Delft-based quantum technology institute Qutech has set out to create a new system to control the qubits in a quantum computer. Though recently built prototype quantum computers can only handle about 50 qubits, the startup – anticipating future developments – is aiming to do up to a thousand.

“We’re developing a modular system that enables our customers to scale up their system by simply adding additional control modules,” explains Qblox cofounder and CEO Niels Bultink. “The module we’re building can control 20 qubits and has dimensions similar to two Monopoly boxes stacked on top of each other. Former systems, using existing electronics, needed a tower of at least two meters high to control the same 20 qubits. If we fill this tower with our modules, we can drive at least 200 qubits.”

The compactness of the system is partly the result of the collaboration with Qutech. Bultink: “In our design, we use technology that follows prior and ongoing developments there. For example, we’ve integrated the functionality of the Qutech waveform generator. This QWG uses an advanced way to produce the waves that are necessary to feed the qubits. It can do so with a very short time delay of several tens of nanoseconds, allowing for the time between a measurement and the subsequent operation to be short compared to the timescale at which qubits can contain their information.”

Qblox system overview
Quantum computer overview. Credit: Qblox

The short time delay is a necessary condition for the quantum computer to be able to restore errors in the system after they’ve been identified, Bultink clarifies. “This avenue, called quantum error correction, is very important as qubits are faulty by nature and an error in a single qubit can screw up the whole calculation. The ability to correct for faulty behavior becomes increasingly important when scaling up. For example, if the probability of an error in one qubit is 10 percent, the probability of an error occurring in a system of three qubits is already increased to 27 percent. For 30 qubits, it even rises to 96 percent.”

By integrating several functions in one device, Qblox has also hugely reduced the amount of cabling. “This again saves space,” says Bultink, “and it diminishes connectivity issues – one of the biggest headaches of quantum computer researchers.”

Decreasing the physical space isn’t the only challenge to overcome in developing a new control stack. To make the system modular, the separate control modules need to work together. This means that they have to be synchronized to send their signals simultaneously. Bultink: “Because the system functions at the nanosecond scale, this is technologically rather difficult. Another challenge is to enable sharing of information between multiple modules – an essential feature for facilitating feedback in the system. The signal measured by module 1 could be an important input for module 18, a few tens of nanoseconds later.”

Qblox rack tower
A control rack filled with five modular systems, together capable of controlling 100-150 quantum bits. Credit: Qblox

Generally applicable

The Qblox system is designed to be compatible with the many different types of quantum computers that are being built now. Some use qubits made up of superconducting structures, others use silicon or diamond platforms. The operational frequency range of the control stack has to be compatible with all of these. Also, the timescale for feeding the qubits can vary for different types of quantum computers, from nanoseconds to milliseconds. “We design our devices in such a way that they can be used for most types,” Bultink points out. “This has some consequences under the hood, but in the end, our system is generally applicable by most customers involved in developing quantum computers.”

The modular control stack proves that Qblox is able to make sophisticated, scalable controllers for the quantum computing industry. The fast-growing company – which started in 2018 with a team of just two and is expected to employ more than 12 people at the end of this year – is eager to continuously improve its products. Bultink concludes: “We want to integrate even more functionality in our devices to make them more versatile, compact and cost efficient. This way, we can provide researchers with a product that enables them to easily scale up their quantum computers and be ready for the future.”