Elon Musk, the visionary yet often contrarian CEO of Tesla, has long argued that lidar—the laser-based sensing technology—is a technological dead end, dismissing it as a “fool’s errand.” Yet, one of his company’s former engineers, Eric Aguilar, sees the matter from an entirely different perspective. Rather than abandoning lidar, Aguilar has made it his mission to refine and revolutionize it. He co-founded Omnitron Sensors, a startup built around the belief that every kind of intelligent machine—whether it takes the form of an autonomous vehicle or a humanoid robot—will ultimately depend on lidar to perceive and understand the physical world with precision and reliability.
Lidar, an acronym for Light Detection and Ranging, operates by sending out pulses of laser light and measuring how long they take to reflect back from surrounding surfaces. This data allows it to construct highly accurate, three-dimensional maps of its environment, identifying objects, measuring distances, and interpreting spatial relationships. Historically, this technology was primarily used in specialized industries such as topographical mapping, surveying, and atmospheric research. In more recent years, however, lidar has evolved from a technical curiosity into a central component of advanced perception systems—especially in self-driving cars, where it acts as a crucial “extra set of eyes.” For example, Waymo’s autonomous robotaxis use lidar to detect road hazards, pedestrians, and obstacles more accurately than cameras alone can manage.
Aguilar co-founded Omnitron Sensors in 2019 and serves as its CEO. His expertise is grounded in more than two decades of experience working directly with sensors and autonomous technologies. Before launching his own company, he contributed to several ambitious engineering ventures, including the cutting-edge drone delivery initiative at Google X, his tenure at Tesla, and his work at Argo AI—Ford and Volkswagen’s joint robotaxi startup, which ceased operations in 2022. His deep familiarity with sensor design and manufacturing makes him particularly aware of lidar’s limitations, as well as its untapped potential.
When Musk announced Tesla’s strategic decision to omit lidar from its self-driving approach, Aguilar—who is trained as an electrical engineer—recognized the rationale behind some of Musk’s objections. Traditional lidar systems are indeed expensive, often bulky, and composed of delicate, moving components that wear down over time. Their manufacturing processes have historically been inefficient and costly, limiting their scalability for widespread use in vehicles. But rather than seeing these flaws as reasons to abandon lidar altogether, Aguilar saw them as engineering challenges waiting to be solved.
Through Omnitron, Aguilar aims to redesign lidar from its very foundation, using silicon and semiconductor fabrication techniques to transform how these sensors are built. His goal is to create a new generation of lidar devices that are both significantly cheaper and dramatically more durable, leveraging the precision and scalability of chip-manufacturing processes already proven in the electronics industry. With this approach, the internal components of lidar systems can be produced in vast quantities—at the level of hundreds or thousands per manufacturing wafer—and with nanometer-scale precision, rather than handcrafted in small batches. Aguilar asserts that this transition to silicon-based microfabrication could make lidar sensors as reliable and affordable as other commodity components in modern vehicles. As he told Business Insider, such innovation could “unlock the market,” persuading automotive giants such as Mercedes-Benz, General Motors, and similar manufacturers to integrate lidar into their next-generation fleets.
The challenges inherent to lidar remain substantial. To observers unfamiliar with its mechanics, lidar might seem almost whimsical—akin to a shimmering disco ball. In essence, the system relies on a laser beam directed at a constantly rotating mirror mounted on an electromagnetic motor. As the mirror spins, the emitted light sweeps across its surroundings, and the reflected signals allow the system to calculate the distance and location of nearby objects. However, improving lidar’s detection range often necessitates a larger, more powerful laser, which in turn demands a proportionally larger mirror. The result is a physically bigger, heavier, and more expensive unit—one that becomes increasingly fragile when placed on a moving platform such as a car. Over time, vibrations, heat fluctuations, and environmental wear can degrade its performance, forcing manufacturers to recalibrate or replace units every few months.
Aguilar notes that the fundamental problem with many existing lidar systems lies in their design philosophy. They rely on numerous moving mechanical components that are both prone to failure and costly to maintain. Automotive manufacturers—known as OEMs (original equipment manufacturers)—are understandably reluctant to install such fragile devices in production vehicles. As Aguilar points out, companies like BMW, Volvo, Mercedes, and GM are unwilling to accept sensors that might need replacement several times a year. His firsthand experience in the industry confirmed this frustration: while working at major automotive firms, he repeatedly observed lidar units breaking down under real-world conditions, necessitating frequent recalibration or replacement.
This is where silicon offers a transformative solution. Silicon-based components can be fabricated as solid-state devices, significantly reducing the number of moving parts. The resulting sensors are lighter, more stable, and capable of enduring temperature variations ranging from the freezing cold of winter roads to the intense heat of asphalt under summer sun. Additionally, because semiconductor manufacturing yields extremely precise and uniform components, the consistency between units is vastly improved. As Aguilar explains, modern chip-making methods allow a single wafer to contain hundreds or even thousands of identical lidar elements, bringing the cost of production down dramatically.
The economic landscape of lidar has shifted remarkably over the past decade. During Aguilar’s time at Google’s moonshot research division, a single top-tier lidar unit could cost around $120,000—a price tag that kept the technology out of all but the most experimental vehicles. Advances in design and manufacturing have since reduced that cost to roughly $10,000 per device. Yet Omnitron’s innovations, according to Aguilar, could push that figure even lower—into the mere hundreds of dollars—making lidar accessible not just for luxury vehicles but potentially for widespread use across consumer car models, robotics, and beyond.
Beyond affordability, Aguilar emphasizes lidar’s essential role in ensuring safety and adaptability, particularly for autonomous vehicles. While cameras and radar contribute valuable information, cameras, in particular, are highly susceptible to poor visibility. They can misinterpret shadows, reflections, or low-light conditions, leading to potential confusion in critical driving scenarios. Aguilar illustrates this with a simple example: an autonomous car approaching a shadow might struggle to determine whether it represents a tangible obstacle or merely an absence of light. Lidar, unaffected by external illumination, can instantly resolve such ambiguity by measuring precise distance data, confirming the true nature of what lies ahead.
The applications extend beyond vehicles. Aguilar envisions a rapidly expanding market for lidar-equipped humanoid robots designed to operate safely and effectively within human environments. Companies such as Agility Robotics are already incorporating lidar into robotic perception systems, enabling machines to sense their surroundings with depth and spatial awareness similar to human vision—yet potentially with greater precision. Just as people use multiple sensory modalities, not only sight but also touch and proprioception, to interact safely with their surroundings, advanced robots will require multidimensional sensing systems. For Aguilar, lidar represents a cornerstone of that sensory architecture, giving robots the ability to estimate distances, detect objects and people, and navigate dynamic spaces dependably.
Ultimately, Aguilar’s argument in favor of lidar resonates with a fundamental principle shared by many leaders in the self-driving and robotics sectors: safety defines success. The standard for integrating autonomous systems into human contexts is—and must remain—exceptionally high. As Aguilar succinctly puts it, people will not entrust their lives or loved ones to machines unless they are convinced those machines are designed with uncompromising robustness. “I’m not going to let this thing hold my baby,” he remarked, underscoring his conviction that dependable sensors like lidar form the backbone of public trust in autonomous technology. By reengineering lidar to be affordable, resilient, and scalable, Aguilar and Omnitron Sensors hope to build that trust—one laser pulse at a time.
Sourse: https://www.businessinsider.com/tesla-engineer-lidar-humanoid-robots-robotaxis-omnitron-sensors-eric-aguilar-2025-10
