Analyzing $VPG's Moat in Humanoids
Tesla’s $TSLA anticipated rollout of toward Optimus Gen 3 production highlights a critical bottleneck that $VPG fills, the nervous system.
I have had a few people as me is their existing foothold for "Customer 1" defensible? After conducting further research, the hardware moat for Vishay Precision Group $VPG in torque sensing is becoming undeniably defensible. Here is why VPG holds a real moat with "Customer 1" aka Tesla Optimus.
1. IP Portfolio and Deep Expertise
Torque sensors are the most difficult sensors in the robot to get right. Unlike tactile sensors in the fingertips, torque sensors in the joints must endure massive heat from the actuators.
The Moat: VPG uses proprietary foil alloys that maintain precision even as the robot’s motors heat up. If a sensor drifts with temperature, the robot loses its balance or crushes the objects it touches.
The Defense: This is not just a patent, but is also 60 years of metallurgical trade secrets. Competitors can’t simply replicate the recipe for these temperature-independent foils.
2. High Switching Costs: The Data Lock-In
Tesla is currently re-tooling its Fremont factory for Optimus production. At this stage, the connective tissue of the robot is locked in.
The Moat: Tesla’s neural networks are trained on the specific physics, noise profiles, and latency of VPG sensors.
The Defense: Switching to a different sensor provider would require Tesla to re-collect a mountain of training data. Ripping out the nervous system now would cause massive delays and model drift, making a pivot away from VPG prohibitively expensive.
3. Reliability and Infant Mortality
VPG has cited long-term reliability as a major foothold vs. the competition. In a humanoid meant for 24/7 labor, a sensor failure is a catastrophic hardware event.
The Moat: The bonding process, how the sensor is molecularly attached to the actuator, is VPG’s specialized IP.
The Defense: Most in-house or startup sensors fail after a few hundred hours due to fatigue or delamination. VPG sensors are transducer class, meaning they are designed to last the entire lifecycle of the robot without losing calibration.
4. The TAM Multiplier
Each Optimus unit requires an estimated 15-20 torque sensors provided by VPG and is generating $850 midpoint revenue per robot in the design phase and at volume production scale (Think 500k+), that number comes down closer to $400 per robot.
If Elon Musk hits his gaudy target of 10M humanoid robots per year by 2030, that would mean ~$4B in humanoid revenue for $VPG. I think everyone knows the biggest question mark for Humanoids is timeline, but it is an eye opening number to think about for a $600M MC company today.
As Tesla moves from prototypes to factory-scale production, VPG is more than a vendor, they are the integrated standard. VPG has been working with "Customer 1" for many years and has been in lock-step with them through the design phase with consistently growing orders. The combination of metallurgical secrets, reliability, and the massive cost of re-training AI models creates a moat that is getting wider by the day.