Google Willow Chip, A Closer Look At The Tech Giant’s Push into Quantum Computing

Google’s latest quantum processor, the Willow chip, marks a significant leap in the quest for practical quantum computing. Unveiled on December 9, 2024, Willow boasts 105 superconducting qubits and has achieved groundbreaking error correction and computational speed advancements. Notably, it completed a complex computation in under five minutes. This task would take today’s fastest supercomputers an estimated 10 septillion years.

Traditional quantum processors generally become less reliable as more qubits are added. Nevertheless, Google AI has changed the game with Willow Chip. They’ve developed a design that supports a high number of qubits. It also improves error correction as this system scales.

In this article, we will explore the history of this revolutionary chip, discuss its innovations and functionality, examine potential applications, and consider what it means for the future of computing.

GOOGLE AI (Alphabet Inc. [GOOG])

Google first entered the quantum race in 2012 with its Quantum AI division, aiming to build a large-scale quantum computer primarily focused on superconducting qubits—tiny circuits that exhibit quantum behavior at extremely low temperatures. The initial goal was to explore how these qubits could outperform traditional computers in specific tasks.

The team collaborated with NASA and the Universities Space Research Association (USRA) to initially access the D-Wave quantum annealer, a specialized quantum system designed for optimization problems. However, Google soon realized that quantum annealing was not the path toward a fully programmable, general-purpose quantum computer. Instead, it focused on developing gate-based quantum processors, which promised greater versatility and computational power.

Hartmut Neven, Google’s Vice President of Engineering and founder of the Quantum Artificial Intelligence lab, emphasized the team’s mission to build quantum computing solutions for unsolvable problems. He highlighted potential applications in areas such as drug discovery, nuclear fusion reactor design, and reducing the energy costs of fertilizer production. Neven also noted the broader implications for quantum machine learning and other fields where quantum effects are significant.

Since then, the company has achieved a series of groundbreaking innovations marking Google’s journey into quantum computing, each building upon the last to push the boundaries of what’s computationally possible.

A Look Back on Google’s Technological Strides

So, before the development of Willow, we will also delve into Google AI’s quantum chips innovations that eventually led to this groundbreaking discovery. From Bristlecone (2018) to Sycamore (2019) and Sycamore 2 (2021), Google has continuously refined its quantum processors, culminating in the Willow Chip (2024)—its most powerful and scalable quantum processor yet.

Bristlecone: Laying the Foundation

In March 2018, Google introduced Bristlecone, a 72-qubit quantum processor designed to test and refine error rates in quantum computations. This superconducting chip aimed to address the challenges of scalability and error correction, which are pivotal for the development of practical quantum computers. While Bristlecone did not achieve quantum supremacy, it provided invaluable insights into qubit coherence and error rates, setting the stage for future advancements.

Sycamore: Achieving Quantum Supremacy

Building on Bristlecone’s foundation, Google unveiled Sycamore in 2019, a 53-qubit processor that marked a significant milestone in quantum computing. Sycamore was designed to perform specific computational tasks at unprecedented speeds. In a landmark experiment, it completed a complex random circuit sampling task in approximately 200 seconds—a feat that Google claimed would take the world’s most advanced classical supercomputers around 10,000 years to replicate.

This achievement was heralded as the first demonstration of quantum supremacy, showcasing the potential of quantum processors to tackle problems beyond the reach of classical machines. However, the Sycamore processor highlighted challenges, such as maintaining qubit coherence and managing error rates during computations.

Sycamore 2: Enhancing Performance and Stability

In 2021, Google introduced Sycamore 2, an evolution of its predecessor with improved qubit fidelity and error correction. While maintaining a similar qubit count, Sycamore 2 focused on enhancing qubits’ stability and coherence times, thereby reducing error rates in quantum operations. These refinements were crucial steps toward building more reliable and scalable quantum systems. Despite these advancements, challenges remained in reducing errors and scaling up the system for more complex applications.

Thus, this is where the Willow Chip, comes in, because unlike these previous quantum processors, which suffered from increasing errors as more qubits were added, this specific quantum chip was designed to be scalable in quantum error despite the high number of qubit count. Which means that for the first time, Google AI has now made it possible for quantum computers to become more stable as they grow in numbers—counteracting the current state of quantum computers!

So, What exactly is the Willow Chip?

For years, quantum computers have been incredibly powerful in theory but impractical in reality, mainly because qubits are notoriously unstable. Even the slightest interference—whether from temperature fluctuations, radiation, or electromagnetic noise—can disrupt their delicate state and introduce calculation errors. The Willow Chip changes that. Instead of suffering from increased error rates as more qubits are added (a problem that has haunted quantum computing for decades), the Willow chip actually reduces errors with more qubits, proving that scalable quantum error correction is possible.

Beyond its stability, the Willow Chip is also equipped with speed. In fact, during their benchmark tests, it solved a computation in under five minutes, which, even in today’s most powerful supercomputer, still takes about 10 septillion years to complete. Pretty amazing, right?

On a deeper level, it has a Scalable Qubit design, meaning that the Willow Chip features 105 superconducting transmon qubits, with each qubit having an average connectivity of 3.47, typically connecting to four neighboring qubits. Additionally, the chip achieves a T1 coherence time (when a qubit remains in a superposition state) of up to 100 microseconds, significantly improving over previous processors.

To maintain their quantum state, these qubits must be cooled to near absolute zero (-273°C) using specialized refrigeration systems. Generally, the main challenge of these quantum chips is qubit instability because the more qubits you add, the more unstable it is, as it is more prone to computational errors. However, Willow’s architecture is designed for error-resistant scalability, meaning that as more qubits are added, the system becomes more stable instead of less reliable, which is thought impossible.

Furthermore, Google’s Willow Chip has already demonstrated computational speeds that are simply unattainable for classical computers. In a recent benchmark test, Willow Chip completed a calculation in under 5 minutes—a task that would take today’s most advanced supercomputers over 10 septillion years.

Lastly, the Willow Chip also features an improved circuit design that allows qubits to communicate more efficiently, resulting in better stability and accuracy in calculations.

Willow Chip’s Competitive Landscape

Looking into the quantum computing race, it is without question that the Willow Chip currently stands out among others in the market. For example, IBM’s Condor was launched in late 2023 as a 1,121-qubit quantum processor, making it one of the most ambitious quantum chips built to date. Condor’s key strength lies in its sheer number of qubits, as IBM aims to push the boundaries of quantum processing power. However, while Condor boasts a high qubit count, it still struggles with error correction, which remains a fundamental challenge in quantum computing.

The next one is Rigetti Computing’s Aspen-M chip, which was launched in 2022. Unlike IBM and Google, which focus on monolithic quantum processors, Rigetti has taken a different approach, meaning its qubits are spread across multiple connected chips. Aspen-M features an 80-qubit setup, making it an important player in mid-scale quantum computing. However, the issue remains with error correction. Not to mention that since the Aspen-M chip has a lower qubit number compared to IBM, its speed and efficiency are quite significantly behind those of IBM offerings.

Lastly, Google’s very own Sycamore, launched in 2019 (54-qubit), is the company’s first attempt to achieve quantum supremacy. This term describes when a quantum computer outperforms even the most powerful classical supercomputers in a specific task. Still, it suffered from significant error rates that limited its real-world usability.