Moon Surgical’s Maestro Brings Physical AI to Surgery

Robotic surgery has long meant a large console, a dedicated room and a price tag that keeps it out of many operating rooms. Moon Surgical, a company based in Paris and San Francisco, is betting on a different shape for the technology: two compact robotic arms that sit at the bedside, hold the laparoscope and a retractor, and let the surgeon keep using the instruments they already know. Over the past eighteen months the company has turned that idea into something larger, an AI platform that now steers the camera, sets itself up and even drafts the operative report.

A robot that assists rather than replaces

Moon Surgical’s product is the Maestro system. According to a peer-reviewed description in Surgical Endoscopy, Maestro is a bedside cart with two arms that grasp standard, off-the-shelf laparoscopic instruments and move in concert with the surgeon, who can manipulate them almost weightlessly. One arm typically manages the scope while the other holds tissue out of the way with haptic feedback. The cart includes a touch interface for the circulating nurse and proximity sensing to help it dock beside the operating table.

That design is deliberate. Rather than building a full telemanipulator, Moon targets the role usually played by a human assistant holding the camera and retracting tissue. The company describes the goal on its own website as an operating room that is “sustainable, efficient, and digitalized.” The FDA cleared the commercial Maestro system in June 2024, as Moon noted in its March 2025 ScoPilot announcement.

ScoPilot: AI that follows the surgeon’s instruments

The feature that put Moon in the AI conversation is ScoPilot. Announced with FDA clearance on March 18, 2025, ScoPilot lets the laparoscope held by Maestro automatically follow a chosen instrument tip, so the surgeon can reframe the view without letting go of the tools in their hands, according to the company release. Moon said ScoPilot runs locally on Maestro using NVIDIA Holoscan and called it the first AI application to run intraoperatively on any commercially available surgical robotic platform with Holoscan, which is a company claim. NVIDIA’s vice president of healthcare, Kimberly Powell, described it in the same release as the first FDA-cleared AI-driven surgical movement natively integrated into a robotics platform and powered by Holoscan.

The Surgical Endoscopy paper explains the mechanics: a deep-learning object detection module finds instrument locations in the laparoscopic video, and a tracking module uses those positions to direct the arm holding the scope. At the time of the study, ScoPilot was trained to track key tools such as the monopolar hook and grasper, with more instruments in development.

A second FDA clearance, announced on July 2, 2025, made that expansion easier. Moon received 510(k) clearance for Wi-Fi and 5G connectivity and for a Predetermined Change Control Plan (PCCP) covering ScoPilot, which the company says lets it build, test and deploy updated versions of the AI more efficiently within a cleared framework. By December 2025, Moon had rolled that connectivity out to all deployed systems, and Maestro had been used to treat more than 2,300 patients, according to MedTech Dive.

Version 2.7 turns Maestro into a multi-model platform

The most recent step came on June 9, 2026, when Moon announced Maestro Software Version 2.7. The release describes three coordinated AI capabilities spanning the whole procedure:

  • Automated Setup: optimizes platform positioning and workspace configuration based on surgeon preferences, room layout, technique and patient anatomy.
  • Enhanced ScoPilot: improved intraoperative assistance trained on newly observed instruments and clinical configurations, enabled by the PCCP.
  • Automated operative reports and coding: the first third-party AI application on Maestro, designed to turn surgical video into post-operative documentation and procedural coding before the surgeon leaves the room.

The update also adds a mobile standby mode for moving the cart between rooms and cuts boot time by 40%, Moon said. The company stated that these models draw on a dataset of more than 3,000 procedures. A follow-up announcement with partner Uncovr reported the first 20 patients whose surgeries were supported by automated operative reporting on the platform.

Evidence from the operating room

Clinical data so far is early but encouraging. The IDEAL 2a study published in Surgical Endoscopy in October 2025 followed 45 “solo” procedures, in which the surgeon operated without a human bedside assistant, performed by five surgeons at three French centers between October 2023 and April 2025. Cholecystectomy, inguinal hernia repair and sleeve gastrectomy were the most common cases. Mean robotic setup time was 6.1 minutes, one case (2.2%) was converted to multiport laparoscopy, and the authors reported no device-related adverse events. The authors note the study was uncontrolled, that two of them hold an indirect equity interest in Moon, and that controlled trials are still to come.

Real-world adoption adds an operational angle. Lee Health’s Gulf Coast Medical Center in Florida reported more than 100 Maestro procedures with no increase in operative time or length of stay, and said the first 25 bariatric cases saved the equivalent of 8.3 full working days of surgical technologist time, according to a July 2025 announcement.

Partnerships widen the reach

Moon is also building an ecosystem around Maestro. In July 2026 it announced a strategic commercial collaboration with KARL STORZ that pairs Maestro with KARL STORZ 3D visualization and instrumentation. The companies framed the deal around “Scalable Surgery,” Moon’s term for making minimally invasive procedures more predictable as surgical volume shifts toward ambulatory and lower-acuity sites. Earlier, Moon expanded its long-running NVIDIA relationship to use the Isaac for Healthcare platform for developing and testing new AI capabilities, including digital twins of the operating room.

What it means

For health systems, Maestro illustrates a shift in surgical robotics from big-ticket consoles toward smaller, software-defined assistants that slot into existing laparoscopic workflows. The most important innovation may be regulatory as much as technical: a PCCP lets an AI feature improve on a planned path rather than waiting for a fresh clearance for every change. Connectivity and third-party apps, meanwhile, hint at the operating room becoming a data platform, where setup, camera control and documentation are all handled by models trained on thousands of cases. Leaders weighing this kind of technology should watch for controlled outcome studies, interoperability with their existing towers and instruments, and clear governance over procedural video and data.

Conclusion

Moon Surgical has moved quickly from a two-arm robotic assistant to a multi-model physical AI platform with FDA clearances, a growing clinical record and partners such as NVIDIA and KARL STORZ. The next test is whether larger studies confirm the workflow gains, and whether automation before, during and after surgery can make minimally invasive procedures easier to scale.

Sources