Erik Lucero, Lead Engineer at Google Quantum AI, recently outlined his team’s ambitions in quantum computing for the years to come. According to the engineer, Google will be able to market a quantum computer before the end of the decade.
In one blog post, Google explained that it wants to build an efficient and useful quantum computer by 2029. To give itself the means to succeed, the company unveiled a new quantum AI campus in Santa Barbara. The facilities house a quantum data center, hardware research labs, and quantum processor chip manufacturing facilities.
As a reminder, quantum computing is a type of calculation based on qubits to encode data instead of the traditional bit, thus allowing the superposition of states. In other words, the processed data could be read in more than one state at a given time (1 and 0 simultaneously). The advantage of this approach is that it exponentially increases the amount of information you can process, and therefore widen our calculation limits.
This is the reason why several companies are investing heavily in these new generation computer techniques. Indeed, it becomes more and more difficult to obtain performance gains with traditional chips.
According to Google, this additional processing power could for example be useful to simulate molecules, and therefore nature, with precision. It could also help design better AIs, more effective drugs, or even more carbon-efficient fertilizers. In short, it would help make the world a better place (at least on paper).
The target announced by Google comes a year and a half after the company declared to have achieved “quantum supremacy”. In other words, one of his computers had for the first time been able to calculate certain operations much faster than a conventional computer.
Sycamore, the name of the processor in question, would indeed have, according to Google, operated with its 54 qubits a very complex calculation operation in just 200 seconds. This was an operation that a conventional computer would have been able to solve in about 10,000 years.
Following the publication of this study, William Oliver, researcher at MIT, had notably underlined the importance of this technical prowess in the field of data processing, comparing it even to the first flight of the Wright brothers in the field of aviation.
“The plane was not the first airborne vehicle to fly and did not solve an urgent transport problem ”, he had explained. “Nor has it announced the beginning of the end of other modes of transport. But we remember this event as proof of a new operational regime, the self-propelled flight of a heavier-than-air plane ”.
Some competitors had nevertheless rushed to cast doubt on the alleged progress of Google, like IBM. According to the engineers of the company, Summit, their supercomputer, would not have needed 10,000 years to perform the calculation proposed by Google, but only two and a half days.