QuEra Computing Automates Critical Quantum Computer Subsystem with AI, Accelerating Commercial-Grade Deployments
A breakthrough in quantum computing has been achieved by QuEra Computing, a leader in the field. The company has successfully automated a critical subsystem of its quantum computers using artificial intelligence (AI), enabling the acceleration of commercial-grade deployments. This development marks a significant milestone for the industry, as it addresses one of the major challenges facing the widespread adoption of quantum computing: the need for human expert intervention to maintain and operate these complex machines.
The AI agent used in this achievement is Anthropic’s Claude, which was developed and validated by QuEra’s team. Claude wrote its own control software, allowing it to recover from laser system failures in seconds, compared to the minutes required by a specialist. This automation not only saves time but also reduces the need for manual intervention, making quantum computing more practical and efficient.
Quantum computers rely on lasers held at precise frequencies to control atomic qubits. However, these lasers drift over time, causing the machine to malfunction unless corrected by an expert. The increased number of lasers in each generation of machines has made this a significant challenge, as every laser requires manual tuning and commissioning. This process can take weeks, tying up scarce human expertise.
QuEra’s team spent two to three weeks writing a recovery script for common laser disturbances by hand. However, this approach had limitations, as it only handled failures that the authors anticipated listing. To overcome these challenges, QuEra turned to Claude and the Model Hardware Standard (MHS), a new standard for AI agents to safely operate physical equipment in scientific research and advanced manufacturing.
The MHS allows AI agents like Claude to run their own experiments on dedicated testbeds, propose fixes, try them out, read results, and refine. This process is continuous, including overnight runs, covering hundreds of failure cases that no team of specialists could work through by hand. Engineers set the scope, reviewed every step, and decided what counted as proof of success.
The software produced by Claude is a conventional, fully inspectable program, not a model making decisions at runtime. Moreover, it operates within safety bounds declared in the standard itself, ensuring safe operation from the outset. This approach has been validated through rigorous testing, with results showing that Claude can recover reliably and quickly, even when faced with real faults.
The pilot project demonstrated several key benefits of this AI-driven automation: reliable recovery, rapid response times, handling of real faults rather than just test cases, improved performance over expert tuning, and the ability to transfer settings between lasers. These achievements have significant implications for the deployment of commercial-grade quantum computers.
Prior to this breakthrough, every laser used in a quantum computer required standing on scarce human expertise time. On-site recovery was necessary at any hour, taking up to half an hour per tuning session, and commissioning a new operating point took weeks. With Claude’s automation, these limitations are significantly reduced, making it possible for more systems to be deployed and properly supported.
Sergio H. Cantu, Vice President of Quantum Systems at QuEra Computing, highlighted the importance of this achievement: ‘For years, the hardest part of scaling quantum computers wasn’t the physics; it was the people driving at 2 am to fix a laser lock.’ He emphasized that QuEra has built a solution using the Model Hardware Standard to address this challenge.
QuEra’s participation in the MHS research preview is an early example of AI partnerships. The company works closely with strategic partners, including Amazon Web Services for cloud delivery and HPE for on-premises high-performance computing integration. This collaboration enables QuEra to leverage cutting-edge technologies like machine learning jobs and apply them to real-world challenges.
The Model Hardware Standard (MHS) is a new standard that allows AI agents to safely operate physical equipment in scientific research and advanced manufacturing. It started as a collaboration between Anthropic and HHMI Janelia Research Campus, with QuEra participating in the limited research preview. The MHS ensures safe operation by declaring bounds, interlocks, and emergency stops within its own standard.
QuEra Computing is at the forefront of neutral-atom quantum computing, helping enterprise innovators leverage this technology to gain competitive advantage. With a public, peer-reviewed path to fault tolerance, QuEra operates globally from Boston, New Mexico, Tokyo, Zurich, and the United Kingdom. As quantum computing moves from ‘one day’ to ‘Day One,’ QuEra delivers practical impact today while leading the way toward large-scale, fault-tolerant systems.
QuEra’s President, Takuya Kitagawa, emphasized that even for experienced companies like his own, maintaining these machines at peak performance is a significant challenge. He highlighted the importance of this breakthrough: ‘A customer expects the entire computer to hold itself together without a specialist in the room.’ This achievement makes it far easier and cheaper to keep QuEra’s computers running at their best.
The results from the MHS research preview, combined with Anthropic’s frontier AI models, are significant. They demonstrate that AI can be used not only for automation but also for improving performance and efficiency. As quantum computing continues to advance, we can expect more innovations like this one to emerge.