Google Quantum AI: New Quantum Chip Outperforms Classical Computers and Breaks Error Correction Threshold
Insider Brief
- Google Quantum AI’s new 105-qubit Willow chip marks a new milestone in quantum computing for the company, demonstrating enhanced computational power, scalable error correction, and a path toward commercially viable systems.
- In a benchmark test, Willow performed a calculation in minutes that would take the fastest classical supercomputer an astronomical timeframe, illustrating the exponential advantage of quantum over classical computation.
- By showing that increasing qubit numbers can actually reduce errors, the Willow chip validates a fundamental approach to quantum error correction, paving the way for large-scale, fault-tolerant quantum machines capable of tackling real-world challenges.
Google Quantum AI announced that it is moving past the Sycamore era and taking another leap down its roadmap with the introduction of the 105-qubit Willow, a new quantum chip that has achieved a milestone in computational power and error correction, marking a major step toward large-scale, commercially viable quantum computing.
The team, which published their findings in Nature, is also eyeing a quantum device that overcomes the limitations of errors and offers real-world solutions to tough problems, the ultimate destination as they progress along their roadmap.
“The mission of the Google quantum AI team is to build quantum computing for otherwise unsolvable problems,” said Hartmut Neven, a vice president of engineering at Google and founder and manager of the Quantum Artificial Intelligence lab, at a recent roundtable about the new milestone. ”So what problems do we have in mind? The first applications will be modeling and understanding systems where quantum effects are important. So that’s the case for common drug discovery, understanding and designing nuclear fusion reactors, bringing down the enormous energy costs of fertilizer production. But it then extends to multiple other areas, such as quantum machine learning.”
Exponential Advantage Over Classical Supercomputers
In a benchmark test using random circuit sampling (RCS), the new chip performed a computation in approximately five minutes that would take the world’s fastest classical supercomputer, Frontier, an estimated 10 septillion years — that’s even longer than the age of the universe. This result underscores the growing divide between quantum and classical computation as quantum systems scale up.
Random circuit sampling is widely used to measure whether a quantum computer can perform calculations beyond the capabilities of classical systems. The latest results, which improve on similar tests performed in 2019 and 2024, demonstrate the exponential speedup quantum processors achieve as they grow in power.
The comparison with Frontier was made conservatively, assuming ideal performance from the classical system. The researchers suggest that the growing divide between classical and quantum means that even as classical computers improve, the gap will continue to widen exponentially.
Error Correction Milestone
Quantum error correction, a cornerstone for reliable quantum computing, also saw a significant breakthrough. Errors, a persistent challenge in quantum systems due to their fragile nature, traditionally grow as the number of qubits increases. However, the new chip demonstrated the opposite: as more qubits were added, error rates decreased exponentially.
“Quantum Information is extremely fragile. It can be disrupted by many things ranging from microscopic material defects to ionizing radiation to cosmic rays,” said Michael Newman, Staff Research Scientist at Google Quantum AI.
Newman said their approach is to implement error correction by making physical qubits work together to correct errors, usually referred to as the creation of a logical qubit.
“The basic idea is you take many physical qubits and you have them work together to represent a single — what we would call — logical qubit,” said Newman. “So, there may be tons of physical qubits, but there might be only three logical qubits, so these qubits all work together to correct errors, and the hope is that as you make these collections larger and larger, there’s more and more and more error correction. And, so, your qubits become more and more accurate. The problem, of course, is that as these things are getting larger, there’s also more opportunities for error and so we need devices that are kind of good enough so that as we make the things larger, the error correction overcomes these extra errors we’re introducing into the system.”
In the study, researchers showed that scaling from a 3×3 grid to 5×5 and then to 7×7 grids of physical qubits reduced errors by a factor of two each time.
This achievement addresses a challenge that has stymied progress since quantum error correction was first proposed in 1994, according to the team. Ultimately, they plan to continue to work on making this system better – as well as implement new algorithmic techniques and innovations – to move them closer and closer to the ultimate goal: a fault tolerant quantum computer.
The results confirm the feasibility of building large-scale quantum systems resilient to errors, paving the way for practical quantum applications.
“Error correction is the end game for quantum computers,” said Julian Kelly, Director of Quantum Hardware at Google Quantum AI. “This is the quantum computer that everyone’s imagined is using, running very large problems and getting interesting applications.”
He added, “Willow was designed with scalable error correction in mind. It’s not just for this demonstration, but this is a technology that can take us into the future.”
State-of-the-Art Fabrication and Metrics
The chip, featuring 105 qubits, was produced in a dedicated superconducting quantum chip fabrication facility in Santa Barbara. Designed for optimal performance, it incorporates error correction into every aspect of its architecture, from gate development to calibration and fabrication.
Its performance extends across multiple benchmarks, reflecting a holistic approach to quantum system design. The chip demonstrates advancements not only in raw computational power but also in system-wide efficiency and adaptability.
Commercial Potential
The chip’s capabilities represent a step closer to achieving quantum systems that can address real-world problems. These systems hold potential for applications in drug discovery, materials science, and renewable energy technologies, including designing efficient electric vehicle batteries and accelerating progress in nuclear fusion. Many of these tasks are currently infeasible on classical systems, awaiting the computational power quantum systems can unlock.
Google is also offering a hint of their commercialization strategy.
While some companies are exploring commercialization and technical improvements simultaneously – for example, selling on-premise units or providing access to the quantum device on the cloud – Google seems to want to perfect the technology, first, and then determine the best way they can make money from quantum.
The team did not specify — wisely — when exactly they expect to hit that quantum advantage inflection point, saying there are challenges ahead.
They write in a company blog post: “At current physical error rates, we might need more than a thousand physical qubits per surface code grid to realize relatively modest encoded error rates of 10-6. Furthermore, all of this was accomplished on a 105-qubit processor; can we achieve the same performance on a 1,000-qubit processor? What about a million-qubit processor? The engineering challenge ahead of us is immense. At the same time, progress has been staggering, and the improvement offered by quantum error correction is exponential. We have seen a 20x increase in encoded performance since last year — how many more 20x steps until we can run large-scale quantum algorithms? Maybe fewer than you think.”
What’s Next
The new chip’s achievements in both computational power and error correction accelerate progress toward the long-term goal of large-scale, error-corrected quantum computers. Researchers are optimistic that such systems will unlock transformative applications across industries. As quantum technology continues to scale, the gap between classical and quantum systems is expected to grow, making this era-defining technology increasingly practical and impactful.
Charina Chou, Director and COO of Google Quantum AI, said: “We are very much interested in offering quantum computing service that can solve real world problems that are not otherwise possible on classical computers.”