QuEra Computing uses Anthropic AI to automate quantum laser tuning

QuEra Computing has successfully integrated Anthropic’s Claude AI to automate the recovery of its quantum laser systems. The lasers, which are essential for controlling atomic qubits, frequently drift and require recalibration to prevent computation halts. Previously, this manual tuning required highly specialized human experts and took between five and ten minutes per instance.

Using the Model Hardware Standard research preview—a framework developed by Anthropic and the Howard Hughes Medical Institute (HHMI) Janelia to allow AI agents to interact with physical hardware—the AI developed and validated its own control logic. In testing, the AI successfully recovered the system in 695 out of 700 trials, achieving a 99.3 percent success rate. Most recoveries were completed in under six seconds, significantly outperforming manual human intervention.

Beyond speed, the AI-driven approach improved signal quality by reducing noise by a factor of five. The resulting software operates as a conventional program with built-in safety limits and emergency stops, allowing engineers to maintain oversight while moving toward more scalable, self-operating quantum computing deployments.