Medical robotics is moving from remote control toward systems that can learn from how clinicians actually perform procedures. Sentante is taking a step in that direction with the European commercial rollout of its CE-marked endovascular robotic platform, initially targeting vascular surgery and interventional radiology. The company says each procedure can generate synchronized force, torque, motion and imaging data that could eventually underpin AI-assisted endovascular intervention.

The next frontier for medical AI may not be another diagnostic model. It may be a robot that can learn how an expert surgeon moves a catheter through a patient’s vascular system.

That is the strategy behind Sentante, a medical robotics company developing a haptic, device-agnostic platform for remote endovascular procedures. Following CE-mark approval in May 2026, the company has begun commercial deployment in Europe, initially focusing on vascular surgery and interventional radiology.

The move is significant because Sentante is positioning its robotic system not simply as a remote-control device, but as the foundation for what it calls endovascular physical AI.

Physical AI refers broadly to AI systems that perceive and act in the physical world. In Sentante’s case, that means combining robotic control, sensors, clinical imaging and AI so a system can eventually assist with — and potentially automate portions of — catheter-based procedures.

The company’s commercial strategy starts with something more immediate: putting the robot into routine clinical workflows and collecting the data generated by those procedures.

Why the first market matters

Sentante’s system is designed for procedures involving catheters and guidewires navigating compliant blood vessels. That makes vascular surgery and interventional radiology logical starting points, spanning procedures such as peripheral arterial intervention, embolization and interventional oncology.

The platform is designed to work with standard catheter and guidewire devices rather than requiring proprietary instruments. Sentante says its system can control up to three devices simultaneously, supporting guidewires from 0.014 to 0.035 inches and catheters from 2Fr to 8Fr.

That device-agnostic approach could be important for adoption. Hospitals already have established device inventories, physician preferences and cath-lab workflows. A robotic platform that requires an entirely new procedural ecosystem would face a considerably higher deployment barrier.

Sentante’s workstation also attempts to preserve the tactile relationship clinicians have with endovascular instruments. Its 1:1 motion-mimicking interface translates the operator’s physical movements into robotic motion while providing haptic feedback.

The company argues that this is important for both clinical usability and AI training. If the robot records how an expert actually manipulates a catheter — including force and torque — the resulting dataset could contain information that a conventional video recording would miss.

Turning procedures into AI training data

This is arguably the most ambitious part of Sentante’s strategy.

Every robotic procedure can potentially generate several synchronized data streams: device kinematics, force and torque measurements and fluoroscopic imaging. Instead of collecting data through a separate research program, the company wants routine clinical procedures themselves to become the source of its AI dataset.

That could create a feedback loop.

More procedures produce more expert demonstrations. More demonstrations can support better models. Better models could eventually provide more useful assistance to clinicians, potentially making the platform more attractive for additional procedures.

The approach resembles developments elsewhere in physical AI, where companies are attempting to gather large datasets of real-world interactions rather than relying exclusively on simulated environments.

But medical robotics introduces a much higher bar. A model that performs imperfectly in a warehouse or autonomous vehicle simulation can be tested repeatedly. A system influencing a medical procedure must operate within regulatory, safety and clinical constraints.

Sentante is therefore starting with data collection rather than claiming autonomous surgery.

A growing evidence base

The company has already accumulated experience across human, cadaver and animal testing.

Sentante reported its first human clinical cases in 2024 in Riga, involving peripheral interventions including balloon angioplasty, stent delivery and embolization.

In 2025, the company demonstrated a transatlantic remote thrombectomy using perfused human cadaver models. It subsequently conducted a live-subject study involving 24 remote thrombectomy procedures from three locations across two continents, according to Sentante.

The company has also integrated NVIDIA Jetson AGX hardware into its latest-generation robot and has developed on NVIDIA Isaac for Healthcare, NVIDIA’s platform for developing and simulating AI-enabled medical robotics.

Those milestones are relevant to Sentante’s AI ambitions, but they should not be confused with evidence that autonomous endovascular surgery is clinically ready. The current commercial product is a robotic assistance and teleoperation platform; the AI roadmap remains an evolving development program.

Why remote intervention still matters

The commercial launch is also happening alongside Sentante’s longer-term push into remote stroke treatment.

Its stroke program has received FDA Breakthrough Device Designation, a regulatory designation intended for certain medical devices addressing serious or life-threatening conditions where preliminary clinical evidence suggests potential for substantial improvement over existing options. The designation itself is not marketing authorization.

The rationale is straightforward: specialist expertise is unevenly distributed geographically.

Sentante says more than 60% of the U.S. population does not have timely access to a thrombectomy-capable center, citing the access challenge behind its remote stroke program.

If remote endovascular intervention can eventually be demonstrated as safe, reliable and economically viable, the implications could extend beyond convenience. A specialist could potentially intervene without being physically present in the same hospital.

That remains a clinical and regulatory challenge rather than an established capability.

The competitive landscape

Sentante is entering a medical robotics market that already includes established robotic systems for vascular intervention.

Siemens Healthineers, Stryker, Corindus/Vascular Robotics and other companies have developed technologies around robotic-assisted cardiovascular or endovascular procedures. Corindus’ CorPath platform, for example, has helped establish the concept of remote robotic vascular intervention in clinical practice.

Sentante’s differentiation lies in combining haptics, device-agnostic control, force and torque sensing and a data-centric AI strategy.

That combination could matter if the company can demonstrate that richer procedural data produces clinically meaningful assistance. It also creates a potential advantage in training AI models because the system captures physical interaction with the device, not merely the resulting fluoroscopic image.

From robot to physical AI platform

The larger significance of Sentante’s rollout is therefore not simply that another surgical robot is reaching the European market.

It is that the company is trying to make the robot itself a data-generation platform.

That strategy will require years of clinical validation, regulatory work and careful AI governance. Force measurements need to translate into reliable insights. Models need to generalize across patients, anatomy, clinicians and devices. And any AI assistance introduced into procedures will need to demonstrate safety before it can become part of routine care.

For now, Sentante is taking a measured route: commercialize the robotic platform in vascular specialties, collect procedure-authentic data and progressively introduce AI assistance as the dataset matures.

If that strategy works, the endovascular robot could evolve from a tool that reproduces a surgeon’s movements into a system that understands the physical dynamics of the procedure — and eventually helps the clinician navigate them.

Market Landscape

Medical robotics is increasingly moving toward a convergence of robotic control, sensor fusion, computer vision, haptics and AI.

Sentante’s model differs from conventional surgical robotics in several respects:

• Device agnostic: Designed around standard guidewires and catheters rather than proprietary consumables.
• Haptic control: Attempts to preserve the tactile feedback clinicians use during catheter navigation.
• Remote operation: Separates the physician workstation from the patient-side robotic system.
• Multimodal data: Combines force, torque, device motion and fluoroscopic imaging.
• AI roadmap: Uses clinical procedures as the foundation for future AI-assisted capabilities.

The broader competitive market includes robotic cardiovascular platforms, image-guided intervention systems and emerging physical-AI medical robotics. The key question will be whether these systems can move beyond mechanical precision to deliver measurable improvements in procedure quality, safety, access and cost.

Top Insights

• Sentante has begun European commercial deployment of its CE-marked endovascular robot, initially targeting vascular surgery and interventional radiology procedures.
• The platform captures force, torque, device motion and fluoroscopy together, creating multimodal procedural data that could support future physical-AI systems.
• Sentante’s device-agnostic design allows clinicians to use standard catheters and guidewires, potentially reducing integration barriers inside established cath-lab workflows.
• The company’s remote stroke program has FDA Breakthrough Device Designation, but designation is not equivalent to FDA marketing authorization or clinical approval.
• Sentante’s strategy depends on turning routine clinical procedures into high-quality training data while introducing AI assistance only as clinical and regulatory evidence develops.

Power Tomorrow’s Intelligence — Build It with TechEdgeAI