On September 17, 2026, Qubot Technology announced the completion of a Pre-A+ financing round of nearly RMB 100 million.

The round was led by Noyu Capital, with participation from Pudong Venture Capital, Pien Tze Huang Yingke Fund, Tongyang Fund and Taihu Aerospace Power Qimingxing. Bitola Capital served as the company’s long-term exclusive financial advisor.

The proceeds will primarily be used to advance regulatory registration, clinical trials and commercialization preparations for Maxwell, Qubot Technology’s fully automated magnetic-control interventional surgical robotic platform.

Qubot’s technical approach combines distal magnetic guidewire control, a robotic advancement mechanism and a VLA (Vision-Language-Action) foundation model for autonomous path planning.

The company is currently focusing on ischemic stroke thrombectomy, a neurointerventional procedure with a narrow treatment window and a high dependence on operator experience.

[Image: MedRobot banner / article cover]

Four Core Modules of the Maxwell System

According to Qubot Technology, the Maxwell platform currently consists of four core modules designed to form a closed-loop system for autonomous neurointerventional procedures.

1. Magnetic-Controlled Guidewire and Catheter

Unlike conventional interventional procedures, in which physicians manipulate guidewires and catheters by pushing and rotating them from the proximal end, Maxwell applies magnetic force to the distal end.

Qubot Technology states that its magnetic-control system can achieve approximately 0.1 mm-level distal-tip control precision.

The company intends this approach to improve control of guidewires and catheters in tortuous vascular anatomy while reducing dependence on the operator’s manual experience.

2. Foton 4D-DSA Digital Twin Module

The Foton 4D-DSA module is designed to provide real-time three-dimensional vascular reconstruction together with positional information for guidewires and catheters.

This information is used to create a digital representation of the vascular anatomy for intraoperative navigation and robotic control.

The Foton module is also one of the products for which Qubot Technology has begun regulatory registration work.

3. VLA Foundation Model: The Decision-Making Core of Maxwell

The third module is the VLA foundation model, or Vision-Language-Action model.

According to the company, the model is intended to perform tasks including lesion identification, three-dimensional vascular reconstruction, path planning and generation of control instructions.

The company is training the model through reinforcement learning in physics-based simulation and world-model environments, with the aim of allowing the system to learn procedural strategies and operating techniques.

For Maxwell, the VLA model is intended to act as the decision-making layer linking image information with robotic execution.

4. Detachable Robotic Advancement System

Maxwell also includes a detachable robotic advancement mechanism.

While magnetic control is responsible for changing the distal direction of the guidewire or catheter, the robotic advancement system controls axial advancement and withdrawal.

The two mechanisms are designed to work together: magnetic control determines direction, while the advancement robot controls forward and backward movement.

Qubot Technology also intends the architecture to allow physicians to operate away from the radiation field.

Industry Reference Points for Magnetic-Controlled Intervention

Magnetic navigation is not a completely new concept in interventional medicine.

U.S.-based Stereotaxis has commercially deployed robotic magnetic navigation technology in cardiac electrophysiology for years. Its Genesis platform uses externally generated magnetic fields to control the direction of magnetically enabled catheters.

In 2026, Stereotaxis’ MAGiC magnetic interventional catheter received FDA clearance, while the company’s magnetic navigation technology has accumulated use across a large number of cardiac procedures.

However, neurovascular intervention presents a different technical environment.

Cerebral vessels can be highly tortuous, the intracranial working space is small, and thrombectomy places strict demands on navigation speed and procedural timing.

Most neurointerventional robotic systems currently under development therefore use mechanical actuation, in which a robot grips and advances conventional guidewires and catheters.

One example is XCath’s Iris system.

Qubot Technology has chosen a different architecture: magnetic distal-tip steering + robotic axial advancement + VLA-based autonomous path planning.

In this comparison, Maxwell differs primarily in its combination of magnetic distal-tip control with AI-based autonomous decision-making.

Stereotaxis has established commercial experience in cardiac electrophysiology, while systems such as XCath Iris use mechanical actuation for neurointerventional procedures.

Qubot Technology is attempting to combine the two control dimensions with an AI decision layer to move further toward autonomous neurointervention.

Clinical Progress: Still in the Regulatory Development Stage

Despite its goal of fully automated intervention, Maxwell remains at the regulatory development stage.

Qubot Technology has initiated regulatory work for both the Foton 4D-DSA module and the complete Maxwell system, covering registration pathways in China, the United States and Europe.

The company plans to conduct animal studies, clinical trials and regulatory submissions progressively between 2026 and 2028.

According to public information, Qubot Technology has also established cooperation with nearly 10 hospitals.

However, detailed information on the nature of these collaborations — including whether they involve early technical cooperation, protocol development or formal clinical trials — has not yet been fully disclosed.

Another issue that remains to be clarified is the eventual regulatory strategy for a system combining magnetic control, 4D imaging and AI-based autonomous decision-making.

Whether the modules will ultimately be registered as one integrated product or proceed through separate regulatory pathways could directly affect the development and commercialization timeline.

Financing and Team

The nearly RMB 100 million Pre-A+ round was led by Noyu Capital, with Pudong Venture Capital, Pien Tze Huang Yingke Fund, Tongyang Fund and Taihu Aerospace Power Qimingxing participating. Bitola Capital served as long-term exclusive financial advisor.

Qubot Technology’s founding team combines backgrounds in artificial intelligence and medical devices.

Joe Wu, Founder and CEO of Qubot Technology, studied mathematics and physics at the National University of Singapore and later pursued graduate studies in computer science at McGill University.

He previously conducted deep-learning research at MILA and Microsoft Research.

Yan Zhou, Partner and President of Marketing, has more than 20 years of experience in the medical device and AI industries and previously held senior positions at multinational healthcare companies.

The company is headquartered in Shanghai and is continuing to build its product development and regulatory capabilities around the Maxwell platform.

Industry Outlook

The commercial logic behind neurointerventional robotics is closely tied to the limited availability of physicians capable of independently performing highly complex procedures such as mechanical thrombectomy.

Many hospitals may have access to interventional imaging equipment but still lack experienced neurointerventional operators.

Qubot Technology’s proposed solution is to reduce this dependence on operator skill through distal magnetic control and increasingly automated robotic execution.

However, achieving fully autonomous thrombectomy will require substantially more than precise guidewire navigation.

The system will need to operate safely across highly variable vascular anatomy, identify and respond to unexpected intraoperative conditions, coordinate multiple guidewires and catheters, and make reliable real-time decisions.

These requirements will require extensive preclinical and clinical validation.

The application of VLA models in surgical robotics also remains at an early stage. Moving from simulation-based learning to reliable autonomous performance in real patients will require clinical evidence demonstrating safety and reproducibility.

Qubot Technology’s current technical route — magnetic distal-tip control + 4D vascular digital twin + VLA autonomous decision-making — provides a technically differentiated approach to neurointerventional robotics.

Following this financing round, the company’s animal study data and progress toward human clinical trials over the next 12–18 months will be important indicators of whether its fully automated approach can advance toward clinical use.

September 19, 2026