It’s been about two months since the robot development landscape changed dramatically in the U.S., making mechanical engineers more vital than ever.

The U.S. Federal Communications Commission (FCC) banned the import of foreign-made “advanced robotic devices” (with some potential Department of War exceptions), blowing up demand for American-made versions.

Specifically, there will be no more import of ground-based robots that weigh more than 4.4 lb/1.9 kg and feature LIDAR (light detection and ranging) sensors or cameras, connectivity of at least 200 kbps, and navigation/perception software.

In other words, small or large AMR or humanoid robots used in the U.S. now must be made in the U.S.

The FCC has also placed new materials restrictions on domestic robots. Now, 65% of the total cost of a robot’s components must come from domestic sources for the robot considered to be “made in the USA.” This will increase to 75% in 2029.

READ MORE: Q&A: RaaS—The Subscription Model That’s Steadily Rewiring Robotics

At the same time these new laws are starting to drive domestic technology development to unprecedented levels, however, American robot makers need to catch up to China in terms of prototype and iteration speed, lowering costs and more.   

From a mechanical engineering perspective, here are some component trends to keep in mind.

1. Sensors Are Improving

The sensing capabilities of humanoid and other robot types continue to advance, with new selections for mechanical and software robotics engineers. Higher-precision LIDAR sensors, depth sensors, ultrasonic sensors and cameras obviously enable better navigation (improved obstacle detection and environment mapping) in dynamic environments.

Looking at LIDAR specifically, which is now common in automated vehicle systems and being integrated into various consumer electronics gadgets as well, there’s a current trend to solid-state technology. Having no moving parts in any component increases reliability and lifespan.

In addition, scientists from NASA and California-based LIDAR tech developer Freedom Photonics explain in a new paper that much smaller and lighter LIDAR systems are being achieved. Some strategies to achieve this use monolithic photonic integrated circuit platforms, hybrid photonic integration and advanced photonic and electronic module packaging.

With regard to improved camera resolution, VGA (a video display controller standard introduced in 1987 by IBM) is still prevalent in current robotic technology, but cutting-edge camera companies are now moving well beyond this. 

Onsemi, for example, is achieving both higher resolution and expanded range depth through using a back-side illumination pixel. These new sensors also provide better quantum efficiency (a measure of a photodetector’s electrical response to light) and reduced depth jitter, enabling better gesture and facial recognition and improved quality control inspection capabilities. 

Another new robotic vision advancement is the very small vision “System-on-Chip” from RealSense, with stereo disparity processing. This platform runs on power-over-ethernet, which the company says help simplify overall robot system design. 

Another new ethernet-powered sensor for humanoid robots has been released recently by Texas Instruments. Its mmWave radar sensor IWR6243 combines camera and radar data to achieve better 3D vision (object detection, localization and tracking) while reducing incidences of false-positive sensor data being sent to the AI software controlling the robot’s movements.

2. Swarm Robotics: Careful Component Choice Required

Swarm robotics, which is a term for many robots moving basically in concert, is an area of interest for some industrial applications, disaster scenarios (removal of debris in earthquakes) and other uses. The robots that will be employed in these applications will likely be humanoid or AMR in shape, requiring high levels of dexterity, high lift-and-carry capacity, excellent vision systems, obstacle detection and more.

As David Chen, author at the Electricalflux.com knowledge base recently explained, when engineers transition from single-agent automation to advanced swarm robotics projects, the fundamental paradigm shifts from centralized control to emergent, decentralized behavior. In a true swarm, “if one robot fails, the collective mission continues uninterrupted,” Chen states. “This requires a radical rethinking of hardware selection, communication protocols, and power management.”

To build a reliable swarm system, Chen notes that engineers must manage inter-node interference, clock drift and distributed sensor fusion, keeping in mind that “choosing the right microcontroller and sensor suite dictates the physical limits of your swarm.”

3. Plastics Advances to Reduce Costs

The costs of humanoid robots were recently discussed at June’s Robotics Summit & Expo in Boston. Panelists from Schaeffler, ASTM International, Boston Dynamics and more were asked if a humanoid robot costing $20,000 to manufacture would ever be achieved. In response, these industry leaders pointed to things like the need for mass production, using sensors that only meet requirements of the task and component standardization.

It’s well-established that wherever plastic can replace metal, savings in cost (and weight, resulting in lower operational power consumption) will result. In August 2026, this approach was officially adopted through a new partnership between two global industry players, materials developer Celanese and VIGOR, a manufacturer of plastic gears and other components. In this project, the companies aim to reduce joint-module weight in humanoid robots by at least 30% through swapping in plastics for metal parts.

READ MORE: Asylon Automates Military Aircraft Inspection with Robots and Drones

Robotics engineers are becoming more adept at combining different types of plastics to optimize costs and performance. Thermosetting plastics (cured to create an irreversible rigid structure to withstand high-stress and/or higher-temperature conditions) such as epoxy resins and fiberglass-reinforced composites are generally used for structural components, but elastomers (rubber-like polymers such as silicone, natural rubber or nitrile rubber which bend or otherwise deform and return to their original shape) are incorporated with hard plastics in subparts requiring flexibility, gripping, shock absorption, etc. 

Elastomer research continues. In a new paper in the journal Nature Communications, a Chinese team demonstrated success in large-strain actuation with a type of artificial muscle that features dielectric elastomers, which are electroactive plastics. 

Conventional dielectric elastomers don’t tend to stay stable when operated at high strain due to mechanical fatigue or electrical breakdown. This team designed one with a bimodal network structure and zwitterionic side groups (molecular structures that contain both anionic and cationic functional groups). They found it demonstrated “a stably-operable strain of 125% at a field of 25 MV/m over 150,000 cycles.” This material, says the team, can also “heal from mechanical or electrical damage.” 

4. Actuator Evolution: Soft Meets Hard?

A recent 2026 McKinsey report on humanoid robot costs puts actuators at 40-60% of the bill of materials; sensing and perception systems at 10-20%; compute and control platforms at 10-15%; and structural components and battery modules each at 5-10%.

Although some humanoid robot developers are using standard actuators or customizing them in-house or in partnership with an external specialist, advancements in new actuators for soft robotics are likely to “spill over” into conventional “hard” robotic systems such as humanoids, potentially offering substantial cost savings down the road.  

Let’s look at a recent review of high-performance soft actuator developments published by a Chinese research team. 

This group points out that these actuators all must offer low power requirements, high load capacity and fast actuation frequency. Current cutting-edge strategies to achieve this can employ one of a wide range of power sources, from electrostatic force, electric current and light to heat or magnetic fields. Here are some specific examples:

Magnet-Driven Actuators

Embedding paramagnetic particles into elastomers creates flexible actuators that can respond to magnetic fields. 

Electrothermal Actuators

When certain new materials are heated, they are able to undergo reversible molecular or lattice rearrangements, producing mechanical deformation. They return to original form once heat is removed.

Flexible Piezoelectric Actuators

When piezoelectric materials are placed between electrodes, and an electric field is applied along the polarization direction of these materials, “they induce mechanical deformation or stress in a specific orientation,” explains the Chinese team. As with electrothermal actuators, “these induced mechanical changes are reversible and disappear when the external electric field is removed.”

Photo-Responsive Actuators

Similarly, certain polymers and liquid crystals can convert light energy into mechanical motion, through reversible deformation. The Chinese team reports that researchers in the Netherlands used this approach in 2020, employing time light patterns in joints to achieve coordinated gaits in a prototype robot.  

Electrostatic Adsorption Actuators

Electrostatic energy can be transferred into mechanical energy (rotating motors) based on electrostatic attraction. No continuous electric current is required for these motors, lowering their power consumption.

5. Global Outlook: Engineering Expertise Needed

To round out this update of component trends in humanoid robots, here’s a quick update about growth in the sector. 

Improvements in sensors, AI, actuators and more enabled humanoid robots to transition out of labs and into real-world pilots and deployments around the world in 2025, as those watching the sector already know. This expansion will continue.

In April 2026, the International Federation of Robotics (IFR) reported on overall global adoption. At that point in time, IFR states that Western European countries were experiencing robot density increase of 3% year-on-year.

READ MORE: The Right Tool for the Job: Practical Advice for Cobots, Physical AI and Robotics Integration

“Eight countries are within the global top 20, which are Germany, Switzerland, The Netherlands, Austria, Italy, Belgium & Luxembourg, France and Spain,” states the IFR. 

The U.S. ranks 8th worldwide with 307 units per 10,000 employees. Canada follows with 241 units and Mexico with 62 units. Asia has an average robot density of 131 units per 10,000 persons employed in manufacturing, an increase of 11%. 

“The economies of the Republic of Korea, Singapore, Japan and Chinese Taipei are among the top ten most-automated worldwide,” added the IFR. “Based on updated labor market data issued by China’s National Bureau of Statistics, China ranks 6th in Asia and 22nd worldwide. It has 166 robots for every 10,000 people employed, which is a year-on-year increase of 17%.”

With the FCC banning imports of AMRs and humanoid robots, continued inroads with humanoid and AMR deployment in the U.S. will depend on growing hardware and software engineering expertise, materials advancements, supply chain stability and other factors such as rate of AI capability increase.

More content from Takeover Week: Automation & Robotics.