With companies like Microsoft, D-Wave and Google working round the clock to make the Quantum Computing dream happen, this has always been relegated to the lab; where researchers expand on the technology, racing to pack as many Qubits into their systems and pushing its computing performance. Even though the technology is still in its infancy, Google wants to push this towards commercial availability, and into the hands of its customers.
Quantum Computing uses atoms packed into something called a 'Qubit'. These atoms, refrigerated to near absolute zero, have multiple states that can be used to calculate complex algorithms. The traditional computer utilizes 1-and-0 states; however, in the quantum world an atom can have both those, and be something in between. This, in conjunction with something called 'quantum entanglement', that can transfer information extremely fast between two interlinked atoms, makes Quantum Computing a very compelling option. However, it still has to prove that it is faster than the average super computer used today; although theoretically, it should be.
Google and others aim to use these computers to speed up its data centers, since these companies believe this new technology will be able to complete certain tasks a million times faster. If you know how the cloud works, this can improve efficiency, freeing up more computing hours and allowing their customers to utilize these systems for increasingly complex calculations. Since Google charges for every minute a customer uses its service, this will also save its customers some money. Quantum computing will also solve the current issue with traditional silicon based processors, where there is a limit to how small, companies like Intel and AMD, can shrink a transistor.
Recently, the technology blog Linus Tech Tips was asked to visit a Quantum Computer in Canada, and in the linked video you can see how massive these things are. Most of the system is relegated to cool down the qubits. Because they are so large, companies like Google will most likely rent them out over the internet; instead of offering them to its enterprise clients for in house deployments.