Discover the

Quantum AI campus

Scroll to begin the journey

Santa Barbara is home to Google's Quantum AI campus, where we bring together our quantum data center, fabrication facility, and cutting-edge research. Take a virtual walk through our headquarters to see where it all happens.

This experience is best with audio

Turn on sound Proceed without sound

Advances in technology for medicine, agriculture, biology, and even computing itself have accelerated thanks to our ability to simulate and predict how things behave in real life. However, we are starting to reach the limit of our power to compute solutions to some problems.

Our team of researchers and engineers at the Quantum AI campus in Santa Barbara are developing highly specialized quantum hardware, software, and algorithms to enable humankind to solve problems that would otherwise be impossible.

We collaborate with people around the world

This is a truly global effort. Our team works with researchers and developers across the globe who contribute to the Cirq library codebase to advance quantum research.

Hartmut Neven on the possibilities of quantum computing

We collaborate with people around the world

This is a truly global effort. Our team works with researchers and developers across the globe who contribute to the Cirq library codebase to advance quantum research.

Doug Strain on software

This is our home base

The Google Quantum AI team calls Santa Barbara home. The campus contains our large-scale research labs, fabrication facilities, quantum data center, and the first quantum computer to ever demonstrate beyond classical computational ability.

Erik Lucero on the Quantum AI campus

Liquid nitrogen trap

We use liquid nitrogen to remove water from the helium gas mixture in our refrigerators. Since nitrogen is a liquid at 77 degrees kelvin (and water freezes at 273 kelvin) we use these thermoses as a cold “trap” for the water.
Cryostat

Each cryostat is cooled down to 10 millikelvin (mK), which makes these locations in our lab some of the coldest places measured in the universe.

This quiet, zen-like environment protects our quantum processors from disturbances like stray magnetic fields and temperature changes that can destroy the quantum information.
Art and inspiration

The art in the lab was created by Forest Stearns, the Quantum AI resident artist. These vibrant murals are inspired by nature's language of quantum mechanics.

Welcome to Google's Quantum AI laboratory

This is our Quantum AI lab, where our team is designing and running experiments to bring an error-corrected quantum computer into existence. Every decision in the lab is intentional, from the choice of materials to the placement of the wires. All around the lab, we work on our fleet of quantum computers in pursuit of our mission.

Sabrina Hong on engineering

Amplifiers

In order to read out the quantum state of our qubits, we have to amplify the signal. From the cold qubit chip all the way to the room-temperature electronics, we boost the signal from the qubits so that we can convert them to a digital signal later.
Wiring

We send signals from our electronics to the quantum processor via coaxial cables. However, not all wires are made equal. Small imperfections can cause big errors in the processor. We chose the material of these wires specifically to deliver high quality waveforms.
Control electronics

Our qubits are controlled with microwaves, which have the same frequency range that cell towers use to communicate with cell phones. Our team developed custom electronics to generate microwave signals that we use to control our qubits and read their state.

Group-work makes the computer work

Our quantum computers have over 10,000 components and the cryostat is just one of them. Today these systems include custom-built electronics, wiring, amplifiers, and a quantum processor.

Marissa Giustina on theory

Shielded qubit chip

The shielding is made of many layers. One layer is made of superconductive aluminum. This layer has a black coating that shields the qubit chip from stray light. The lid also has a layer made of a nickel-iron alloy that shields the chip from magnetic fields.
Coax cable connector

Each coax cable connector is routed to a qubit. The connectors bring a microwave signal from the control electronics to the chip.
Printed circuit board

A circuit board gives the qubit chip a firm place to attach to the other components of the computer. The qubit chip sits on an aluminum-plated board, printed with connections from the chip to a microwave connector. The microwave connectors attach to the signal wires.

This is a packaged quantum processor

The quantum processor is similar to a CPU that sits on a motherboard. Our quantum processor is a highly specialized chip bonded to a circuit board. The board provides support and connections for the chip to interact with other electronics. The chip is covered by layers of protective shielding.

Marissa Giustina on building quantum hardware

Signal lines

We etch metal wires on a silicon wafer to carry signals to and from our qubits. These lines enable us to control each of our qubits.
Readout resonators

Some of our signal lines connect to our readout resonators. To read the state of our qubits, we activate a resonator with a controlled signal and measure the feedback.

The quantum chip: wiring layer

The wiring chip is where we route signals to and from qubits. We keep this layer separate from the qubit layer to minimize the interaction between the signals and the qubits. The wiring chip is a surface where we have etched wires to send and receive signals to and from the qubits.

Andrew Dunsworth on fabrication

Qubits

Our qubits are remarkably simple: they're just inductors and capacitors. We etch these components onto an aluminum-coated silicon wafer, and provide connections that enable us to control each of the qubits individually.
Indium bump bonds

Our qubit chip and our wiring chip are glued together using indium. These bumps of indium are 10 microns on a side and they also conduct signals.

The quantum chip: qubit layer

The qubit chip is where we bring together the world of quantum mechanics with modern manufacturing. The qubit layer contains the superconducting metals and magnetometers that make up our qubits.

Andrew Dunsworth on fabrication

Qubit

This image is made by plotting data that we collected from our scanning electron microscope. You can see some of the components of the qubit.

The qubit

A qubit is a very sensitive electromagnetic detector made of an inductor and a capacitor, which are wired in parallel. The inductance comes from the Josephson Junction, a sandwich of superconducting metal and an insulator; the capacitance comes from the two metal plates in the shape of the letter X.

Hartmut Neven on the possibilities of quantum computing

Chief scientist

After we demonstrated that beyond-classical computing is possible, we took some time to think ahead. We have a very talented team who has worked hard on a plan to build an error-corrected quantum computer.

In the spirit of John F. Kennedy speaking about landing Americans on the moon, we think we can do it before the decade ends. In practice, that means we think we can build a quantum computer with a million physical qubits by then.

Our next big milestone is a demonstration of a logical qubit or rather reduced logical error. Error correction works by introducing redundancy, so you have to take the information encoded in your logical qubit and distribute it over a set of physical qubits.

If everything goes well and you have a large suppression factor, your device should essentially stay coherent until you switch off your computer.

Eventually, you will have two logical qubits at hand, and you have a full gate set between them. At that point, you really have what may be called the integrated circuit of quantum computing. You have a tileable digital module.

To build a large machine, you would just take this module and replicate it. Hopefully, this is not too far out. Past that point, the risk profile for quantum computing comes way down. It becomes more the risk profile of building a high rise or freeway, rather than doing cutting-edge quantum electronics.

So we have a tight sequence of difficult but reachable milestones that eventually take us to this end goal.

What would we do with this machine? What are attractive, commercially relevant, valuable, or scientifically interesting algorithms to run? There's definitely some more ingenuity needed here, but it's ripe for doing something interesting.

Doug Strain on software for quantum computing

Quantum computing software lead

Our software team connects the different pieces of quantum computing together using open source software and Cloud APIs. Everyone from our external partners to our internal researchers use our software.

Our mission is to make quantum computing accessible and useful to both high-powered researchers, as well as the aspiring programmers, who are just learning about quantum computing right now.

Our products include open-source libraries like Cirq, OpenFermion, and TensorFlow Quantum. Cirq is our canonical way of defining and modifying quantum circuits. With Cirq, you can design and create a quantum circuit, analyze it using simulators, and then send it to hardware using our quantum computing service API. Similarly, OpenFermion is a library for quantum simulations, particularly quantum chemistry and electronic structure calculations.

No one really knows the best way to program code for a quantum computer yet. The world is still developing algorithms, tricks, and methodologies to get the most out of these novel devices. It is our job to distill these tricks people are learning right now and make them reusable so that everyone can try them.

We need to balance the complex tuning needed to squeeze the most out of these complex devices with the usability and flexibility that people expect from modern development tools. And we need to make it understandable so that we can train a new generation of quantum programmers for tomorrow's devices.

As quantum computers get more powerful, the circuits and programs they can execute will get much bigger, both in the number of qubits as well as their length and complexity. We need to advance the state of quantum computers from the 1960s-era of machine code in Assembly Language, where it is now, to the complex abstractions that will be needed as we approach giant error-corrected devices of a million or more qubits.

Erik Lucero welcomes you to the Quantum AI campus

Lead engineer, quantum operations and site lead Google Santa Barbara

Production quantum hardware requires system integration and the hardening of all of our proven research. It also requires us to demonstrate that we can make systems more and more reliable with each subsequent repetition.

What are the challenges along the way that we need to address?

First, we have to assemble an awesome team. This is ongoing. I'm helping to build that team everyday. We have a great starting point and we will continue to grow our team in new dimensions.

Second, we’ll need to plot the course and be delighted to wayfind along the way. We have a hardware roadmap to an error-corrected quantum computer. That is the course. And as with any planned adventure in nature, something exciting and challenging will likely pop up and force us to re-plot that course.

And third, we’ll continue to build the landmark Quantum AI campus. Our error-corrected quantum computer will be the size of a tennis court. It will be an architectural and scientific wonder within which a team of 1 million qubits will operate in concert directed by a surface code error correction, while pumps and cryogenics cheer along.

I invite you to enjoy a peek inside to learn more about our team and our passion for our work. I hope that you're as inspired as we are to embark on this journey to build an error-corrected quantum computer for the world from wherever you may call home.

Sabrina Hong on engineering quantum computers

Quantum systems engineer

As the team works toward a error-corrected quantum computer, part of my job is to develop the experiments and software that can get us from a fabricated chip in a package to a programmable quantum computer capable of error correction.

In particular, I work with the team of folks developing techniques and infrastructure for calibrating, characterizing and maintaining high fidelity operations on a quantum computer.

When the fabrication team wants to test a new strategy, or the design team wants to test a new type of layout, my work in device calibration and characterization gives them the feedback they need to determine whether or not to continue down that path.

To run quantum algorithms, we have to schedule and perform an intricate sequence of precise operations like microwave pulses and fast EC pulses. Part of my job is to develop and study different experiments that can reliably calibrate these operations.

As we start to scale up to larger and larger processors, we need to ensure that this process can still be done quickly at the scale of thousands and even millions of qubits. This means automated lightweight testing that can probe and correct the quality of our gates in the presence of any drifts without interfering with the operation of the computer itself.

Marissa Giustina on the theory of quantum computing

Research scientist

I build cryogenic hardware that facilitates information transfer between the messy and warm world that we live in (at room temperature), and the quantum processors that live in the 10 millikelvin world at the bottom of the dilution refrigerator.

If you look at our system as a whole, you will see racks of electronics. Those are electronics operating at room temperature and they're connected to a big cylinder, which is a dilution refrigerator. Our refrigerator is warmest at the top and it gets colder as you go down. When the fridge is open you'll see that those blue cables at the top connect into some silver cables inside the fridge, and so on down to the bottom. Eventually you reach the package and inside of that package is the processor itself.

A quantum computer is a big system with a lot of supporting infrastructure—thousands of components—but the components of the system keep changing, and the design of each component itself is its own research topic. Yet they have to be developed in a way where they can integrate with each other.

The system as a whole is too complicated for any one person to be able to understand and orchestrate it all. Building the system requires a lot of research, a lot of clear communication, a lot of carefully coordinated creativity.

Andrew Dunsworth on fabricating quantum chips

Quantum processor designer

You have to consider three things when you design a chip:

Planning the physical layout of the processors: our processors are something like very tiny radios that work at very low temperatures.
Fabricating the processors: this uses industry-standard microfabrication techniques.
Testing the processors: we cool the processors down to millikelvin temperatures then shoot microwaves at them, perform algorithms, and take key readings like quantum coherence.

To understand how to design a processor better, you have to understand both the limitations of fabrication as well as what things to measure about the processor. Based on your understanding of how those things are measured, you can start to improve the processor performance as a whole.

I collaborate closely with teams like the packaging team, which is involved in making sure that the signals are properly isolated before they even get onto the chip; or the calibration team, which characterizes and comes up with new ways to calibrate the system. We take the data from the calibration team and feed it back into the design to try to improve things further.

Running useful algorithms on a quantum processor requires many coherent operations. Currently, our processors can't stay coherent long enough to perform long computations.

To overcome that, we have a strategy where we do something called error correction. Error correction is the idea that you can distill a very good qubit out of many not-so-great qubits.

However, error correction comes at a cost. For the strategy to work, you have to use a lot of physical qubits to have an acceptable error rate for a single "logical" qubit.

The major challenge that we're trying to overcome is to better control or isolate our qubits. Isolation is hard because for the qubit to be useful, we need to maintain a high-degree of connectedness between qubits so that they can still perform physical operations with high fidelity.

Our mission is to advance the state of the art of quantum computing and develop the tools to operate beyond classical capabilities, hoping to enable humankind to solve problems that would otherwise be impossible.

Discover the Quantum AI campus Scroll to begin the journey

Santa Barbara is home to Google's Quantum AI campus, where we bring together our quantum data center, fabrication facility, and cutting-edge research. Take a virtual walk through our headquarters to see where it all happens.

We collaborate with people around the world

Hartmut Neven on the possibilities of quantum computing

We collaborate with people around the world

Doug Strain on software

This is our home base

Erik Lucero on the Quantum AI campus

Liquid nitrogen trap

Cryostat

Art and inspiration

Welcome to Google's Quantum AI laboratory

Sabrina Hong on engineering

Amplifiers

Wiring

Control electronics

Group-work makes the computer work

Marissa Giustina on theory

Shielded qubit chip

Coax cable connector

Printed circuit board

This is a packaged quantum processor

Marissa Giustina on building quantum hardware

Signal lines

Readout resonators

The quantum chip: wiring layer

Andrew Dunsworth on fabrication

Qubits

Indium bump bonds

The quantum chip: qubit layer

Andrew Dunsworth on fabrication

Qubit

The qubit

Our mission is to advance the state of the art of quantum computing and develop the tools to operate beyond classical capabilities, hoping to enable humankind to solve problems that would otherwise be impossible.

Our efforts to advance quantum computing for everyone begin here.

Discover the

Quantum AI campus

Scroll to begin the journey