NVIDIA, Disney, DeepMind release 'Newton' physics engine

PLUS: China's surgical robot rally, Ericsson hires humanoids, Restoring the surgeon's touch


NVIDIA, Disney, DeepMind release 'Newton' physics engine

Welcome back to your Robot Briefing

Three tech giants just joined forces on a problem that's kept robots stuck in warehouses: NVIDIA, Disney Research, and Google DeepMind built Newton, an open-source physics engine that trains robots to walk through sand, gravel, and other unpredictable terrain that causes expensive failures in real deployments. Disney already used it to get their BDX Droid moving through shifting surfaces, and now any company can access the same simulation tools.

The question for businesses evaluating automation: does this infrastructure breakthrough finally make outdoor robots viable for construction sites and farms, or are we still years away from reliable deployment?

In today's Robot update:

NVIDIA, Disney, and DeepMind release Newton physics engine
China's surgical robot pricing guidelines spark market rally
Ericsson deploys humanoids at Texas experience center
Soft robotic fingertips restore surgeons' sense of touch
News

NVIDIA, Disney & DeepMind drop 'Newton' physics engine

NVIDIA, Disney & DeepMind drop 'Newton' physics engine

Image Source: Gemini / There's A Robot For That

Snapshot: NVIDIA, Disney Research, and Google DeepMind released Newton, an open-source physics simulation engine that teaches robots to navigate unpredictable terrain like sand and gravel — addressing one of the biggest barriers to deploying robots outside controlled factory floors.

Breakdown:

Disney already deployed Newton to train its BDX Droid to move through shifting sand and gravel with precision, demonstrating real-world application beyond lab testing.
The collaboration between three major players signals that simulating complex physical interactions has become critical infrastructure for the robotics industry, not just a research problem.
Newton specifically tackles granular materials and deformable surfaces, the types of environments that cause expensive failures when robots trained in pristine simulations meet messy reality.

Takeaway: This moves the industry closer to robots that can handle unpredictable environments without constant human intervention, which directly impacts when automation becomes viable for industries like construction, agriculture, and outdoor logistics. The fact that it's open-source means smaller companies can now access simulation capabilities that previously required massive R&D budgets.

News

China's new pricing playbook sparks surgical robot rally

China's new pricing playbook sparks surgical robot rally

Image Source: Gemini / There's A Robot For That

Snapshot: China's National Healthcare Security Administration released the country's first national pricing guidelines for robotic surgery on January 20, establishing a three-tier fee structure that sent surgical robot stocks soaring by double digits.

Breakdown:

The new framework creates three pricing tiers based on robot involvement: navigation (basic guidance), participatory execution (shared control), and precision execution (full robotic control), with higher fees for procedures requiring greater robotic precision rather than charging based on brands or surgical approach.
Market response validates the significance: leading companies saw immediate gains with Tianzhigang up 12%, MicroPort Medical Robot up 14%, and Jingfeng Medical up 9% as investors recognized this removes a major commercialization barrier that had kept hospital adoption inconsistent across provinces.
The pricing shift accelerates the business model transition from one-time robot sales to recurring revenue, as Tinavi Medical Technologies (which performed over 35,000 of China's 48,500 robotic procedures in the first three quarters of 2025) expects consumables and services to become its primary growth driver as procedure volumes increase.

Takeaway: This marks China's surgical robotics market moving from experimental to mainstream, solving the pricing uncertainty that kept hospitals cautious about procurement. The framework's focus on clinical value over hardware costs signals that regulators are ready to scale adoption, making this a template other markets may follow as they grapple with similar commercialization challenges.

News

Ericsson hires humanoids for its Texas experience center

Snapshot: Telecommunications giant Ericsson is deploying Realbotix humanoid robots at its Imagine Studio in Plano, Texas, to handle workforce training and visitor engagement. This marks one of the first public deployments of humanoid robots by a major enterprise for operational functions beyond manufacturing floors.

Breakdown:

The robots will serve three core functions : delivering personalized employee training and career development pathways, greeting visitors with immersive demonstrations of conversational AI capabilities, and supporting campus outreach to students exploring tech careers.
Realbotix's proprietary vision technology enables the robots to autonomously detect movement and emotions, remember faces, recognize colors, and read text, creating more natural interactions than traditional kiosks or displays.
The deployment signals growing enterprise confidence in humanoid robotics as practical workplace tools rather than experimental novelties, with Realbotix CEO Andrew Kiguel positioning this as validation that humanoids can function as partners in customer-facing and internal operations.

Takeaway: Major telecoms don't typically adopt unproven technology for customer-facing environments, making this a credible signal that humanoid robots are ready for enterprise deployment in controlled settings like experience centers and training facilities. Companies evaluating visitor engagement or workforce development tools should watch how this deployment performs over the next 6-12 months as a practical test case.

News

Restoring the surgeon's touch with soft robotics

Snapshot: European researchers are building soft robotic "fingertips" that could give surgeons their sense of touch back during minimally invasive and robotic operations, with the first prototype expected by March 2026.

Breakdown:

The EU-funded PALPABLE project addresses a critical trade-off in modern surgery: while keyhole and robotic procedures reduce patient trauma and recovery time, they eliminate the tactile feedback surgeons need to detect tumor margins and tissue abnormalities during operations.
The probe embeds hair-thin fiber-optic cables in a soft silicone dome that translates physical pressure into visual stiffness maps, allowing surgeons to "see" what they used to feel when pressing directly on organs and tissue.
A European consortium including Queen Mary University of London, Fraunhofer Institute, and Hellenic Mediterranean University is developing the technology, with validation in lab tests before patient use and an expected prototype in March 2026 .

Takeaway: This matters because robotic surgery continues to expand in operating rooms, but surgeons currently operate with significantly reduced sensory information compared to traditional open surgery. The near-term timeline and credible research consortium suggest tactile feedback solutions could become standard features in surgical robotics within the next few years, not a distant possibility.

Other Top Robot Stories

XPeng completed its first ET1 humanoid robot unit built to automotive-grade production standards, with CEO He Xiaopeng confirming the milestone as a critical step toward mass production of advanced humanoid robots by the end of 2026—signaling that major EV manufacturers are transitioning humanoid development from prototypes to production-ready systems.

UBTECH deployed its Walker S2 humanoid at SANY RE's 5G-enabled wind power smart factory in China, marking the first humanoid robotic worker in the renewable energy manufacturing sector where the robot handles precision sorting and adaptive manipulation tasks in a fully operational production environment.

🤖 Your robotics thought for today:
NVIDIA, Disney, and DeepMind just open-sourced Newton so any startup can train robots on sand and gravel without a massive R&D budget—so why are construction and agriculture companies still waiting for "proven solutions" instead of spinning up their own pilots this quarter?

Until tomorrow,
Uli

NVIDIA, Disney, DeepMind release 'Newton' physics engine

Great! Check your inbox and click the link to confirm.
Please enter a valid email address.