Twice, the international research community has by now demonstrated that quantum computers are able to solve specific computing tasks much faster than conventional supercomputers: In 2019, internet giant Google put forward its quantum processor Sycamore that’s based on 53 cryogenic supraconductive qubits (see info box). In record time, the microchip that was cooled down to near-absolute zero of minus 273.135 °C on the centigrade scale (minus 459.67 °F) to ensure reliable operation determined the probability with which the coupled qubits produce specific exotic sequences of numbers in the readout.

A little more than a year after Sycamore, another computing record was set: At the end of 2020, the quantum processor prototype developed by a team of researchers led by Jian-Wei Pan at the University of Science and Technology of China solved a computing task for which a classic computer would need 2.5 billion years – thereby showing so-called quantum supremacy as well.

Independent observers feel that the research work about which quantum researcher Pan reported in “Science” magazine is convincing: Jian-Wei Pan, a former PhD student of Austrian quantum pioneer Anton Zeilinger, has long been regarded as a leading expert in experiments with light particles with which he also demonstrated quantum supremacy. Like IBM publicly challenged Google’s results in 2019, Google expressed doubts about whether classic computers would really do so poorly in solving the Chinese team’s computing task. The task was to predict the passage of photons in a highly complex labyrinth of beam splitters, mirrors and prisms. Pan’s optical apparatus used 76 qubits – more than Google’s 53 – and that may have been one of the reasons for Google’s slight chagrin.

Physical limits for quantum computers

However, in spite of all the progress that’s been made, both Google’s Sycamore and Pan’s quantum processor are still completely unsuitable for practical applications like calculating chemical structures in the fields of medicine and material science, route optimization for autonomous driving, or portfolio analyses. “To be able to compute truly useful tasks with a quantum processor I need somewhere in the neighborhood of a thousand or better yet several thousand qubits,” says Simon Baier, a quantum physicist at the University of Innsbruck. “When you look at the cryostats of Google’s or IBM’s quantum computers and see how many countless cables are already running into them it becomes clear that at some point there’s no more room for even more cables.” In other words, Baier expects the implementation of a quantum computer to entail physical limits clearly before the level of a thousand qubits can be attained. This problem might be solved, though, by interlinking individual quantum processors in a quantum computing cluster.

Baier, who was doing research at the Dutch quantum research center QuTech in recent years, is working on a technology that is to enable this in the future: the so-called quantum internet. Together with former QuTech colleagues, the scientist described the fundamentals of such a network in a pioneering work published in “Science” magazine this April. In the laboratory, the team managed to connect three separate quantum processors in the world’s first multi-node quantum network. In previous attempts, only two quantum devices were successfully interlinked. Diamonds with so-called nitrogen vacancies serve as a minimal version of the quantum memory required for such a network; the communication between them is handled by photons exchanged via fiber optic cables.

Quantum internet achievable in seven years

Even this first rudimentary quantum network is still far from practical implementation – which was not the objective of the work anyway. But it does demonstrate that such networks will be feasible in the future. The architecture the team selected is extensible, says Baier: “Basically speaking, we can add any number of nodes to the network, practically by means of copy-and-paste.” So, the problem of the quantum internet’s scalability has been solved. Even so, several years will more than likely pass before a usable technology is available. A lot of development work still has to be done to enable the transportation of information across long distances. “For the photons from the quantum memory to have the suitable wavelength, a conversion step between the memory and fiber optics is required,” says Baier.

Another question yet to be answered is which protocol language a quantum internet should be based on. Baier expects this hurdle to still be overcome by the EU’s ongoing flagship pro­ject that’s funded with one billion euros, in ­other words within the next seven years. That’s when quantum networks for secure communications for banks and government authorities will be possible, as well as the design of telescopes and atomic clocks with even higher precision than those existing today, and quantum computer networks with higher computing power. For the latter, though, the quantum internet is not an absolute requirement. Andreas Wallraff from ETH Zurich, for instance, is working on interlinking individual quantum processor cores to form a kind of cluster or local area network (LAN) by means of a permanent “quantum connection.” At the beginning of 2020, the researcher’s team established the so far longest microwave-based quantum connection in the world with a length of five meters (16 feet) at the Quantum Device Lab of ETH Zurich.

Faster computing in the quantum LAN

The ETH physicists were able to demonstrate not only that their quantum connection can be cooled adequately – it has to have the same cryogenic temperature as the quantum processing cores – but also that quantum information can actually be transmitted with it reliably. “This does mark a milestone achievement for us,” says Wallraff, “because with it we can show that quantum LANs are basically possible.” Since, like Baier, he expects that the number of qubits of an individual quantum computer cannot be increased to any desired number, Wallraff estimates that quantum computers will depend on such quantum connections in the next 10 to 20 years. His team has already begun to work on longer ones: for the current 30-meter (98 feet) version, a dedicated room was prepared at ETH.

IBM wants to resolve the space issue for the qubits and the technology they require in the current cryostats in a different way: a super-fridge named “Goldeneye” is supposed to keep the quantum processors that will become increasingly complex in the future at the right temperature. The three meter-tall (10 ft) and two-meter-wide (6 ft) behemoth has been designed “with a million-qubit system in mind – and has already begun feasibility tests,” writes Jay Gambetta, IBM Fellow and Vice President IBM Quantum, in IBM’s research blog. “Ultimately,” the IBM man continues, “we envision a future where quantum interconnects link dilution refrigerators each holding a million qubits like the intranet links supercomputing processors, creating a massively parallel quantum computer capable of changing the world.” The goal of a fault-tolerant quantum computer, Gambetta feels, is achievable within the coming decade.