Today, Google’s Quantum AI team unveiled Willow, a state-of-the-art quantum computing chip that has demonstrated the ability to not only exponentially correct errors, but also process certain computations faster than supercomputers could within known timescales in physics.
This is a significant milestone in the Quantum AI team’s journey to create a reliable quantum computer that can expand human knowledge for the benefit of all people. Quantum is a new approach to computing, where people are building machines that use quantum mechanics — the fundamental language of the universe — to break through the limits of classical computing.
Step inside the Google Quantum AI lab to learn more about how quantum computing works and understand six key quantum concepts.
1. Quantum computing: why everything else is “classical computing”
Quantum computing is an entirely new style of computing. Most people are familiar with classical computing: the binary digits (or “bits”) that can be either 1’s or 0’s, which power everything from graphing calculators to massive data centers, and underlie almost all of the digital innovation from the past half-century.
Quantum computing is different. Rather than using classical bits, quantum computing uses quantum bits, or “qubits.”
2. Qubits: the building blocks of quantum computing
Qubits behave according to the laws of quantum physics. Instead of being confined to the “either/or” of binary 1’s and 0’s, they can exist as a blend of both. Qubits can store information in superposition (multiple states at the same time) of 0 and 1. They can also be entangled with each other to make even more complex combos — e.g., two qubits can be in a blend of 00, 01, 10 and 11. When you entangle lots of qubits together, you open up a vast number of states they can be in, which gives you lots of computational power. Those two special properties provide quantum computers with the superpower to solve some of the most difficult problems much, much faster than regular, classical computers can.
3. Fabrication: how the Quantum AI team makes chips for qubits
Unlike classical computing chips — which are produced by a huge and well-established industry — quantum is such a new style of computing that Google makes our own qubits in-house with superconducting integrated circuits. By patterning superconducting metals in a new way, we form circuits with capacitance (the ability to store energy in electrical fields) and inductance (the ability to store energy in magnetic fields), along with special nonlinear elements called Josephson junctions. By carefully choosing materials and dialing in the fabrication processes, we can build chips with high-quality qubits that can be controlled and integrated into large, complex devices.
4. Noise: building packaging to protect quantum computers from disturbances
Quantum computers can be prima donnas. They have the ability to solve problems that would be impossible on classical computers, but they’re also highly susceptible to errors from “noise,” or disturbances like radio waves, electromagnetic fields and heat (even cosmic rays!). So — much like building a sound studio for recording artists — to protect the integrity of quantum computing processes, the Quantum AI team builds special packaging to reduce the noise. They place qubits in this special packaging to connect them to the external world while shielding them from external disturbances as much as possible. Achieving this requires extensive and highly complex mechanical and electromagnetic engineering work, as well as a focus on details such as choosing the right materials or deciding the specific locations to put holes for circuitry.
5. Wiring: creating the pathways to control a quantum computer
Controlling a quantum computer requires sending signals through environments with temperatures of extreme variations. We control qubits with microwave signals, which are delivered through special wires from room temperature all the way to extremely low temperatures. Those wires are chosen to ensure we can deliver signals in the most efficient and accurate way possible. Adding elements such as filtering in the middle of those wires further protects our qubits from being affected by external noise.
6. Dilution fridge: one of the coldest places in the universe
Operating superconducting qubits requires us to keep them at extremely low temperatures that are colder than outer space. A special piece of equipment called a dilution fridge is needed to reach these ultra-cold and dark conditions. By keeping our qubits inside the dilution fridge, the superconducting metals can enter their zero-resistance state — a frigid state where electricity can flow without energy loss — and we can reduce unwanted things like thermal noise. In this way, our superconducting qubits can maintain their quantum properties and perform complex calculations for quantum computing.
Willow is the latest step in our Quantum AI team’s work to unlock the full potential of quantum computing. Now that you’ve gotten a sense of our lab work, check out our quantum computing roadmap to see how we’re planning to bring quantum out of the lab and into useful applications.