The Case for Humanoid Robots

Working in the robotics industry, I’ve come across a lot of new opinions — humanoid robotics and its recent surge in number of startups and media coverage has stirred a lot of conversation. As a result, I decided to synthesize a few ideas as a response to some of the common criticisms.

Frequent Comments

It’s become a trend to bash on humanoid robots. Some of the primary objections are rooted in practical concerns: why design humanoid robots when specialized automated systems already excel at many of these tasks? For instance, we have dishwashers that outperform any dish-washing humanoid, and Roombas that efficiently clean floors without human-like limbs. The humanoid form isn’t naturally optimized for many of these modern tasks; instead, creating robots with forms suited to specific functions often seems more effective.

Another common critique focuses on the current state of robotics research. Companies like Figure have produced humanoids that are often shaky and struggle to perform even basic tasks like standing or walking. Many of these robots are far from autonomous, requiring tightly controlled environments, as seen with Tesla’s Optimus or GXO’s robots working in highly managed warehouse settings. Why not just put wheels on these robots or design them as quadrupeds to enhance stability and functionality? It’s understandable why skeptics question the practicality and value of pursuing humanoid forms for robotics.

These are valid concerns. Brushing off these criticisms only fuels a larger Us vs. Them mentality and neglects the nuanced arguments against humanoid robots. Similarly, short-form rebuttals, like those often seen on Twitter, don’t offer enough space to present a compelling counterargument.

To meaningfully address these critiques, it’s useful to consider the robotics industry from an investor’s perspective and examine why people might want to be optimistic about robotics’ potential, enough to put money on it now. Too many people have restated the same “the world is made for the human form” argument — let’s have more than one reason.

So Why Humanoids?

Let’s address the first comment: humanoid robots is that they are unnecessary because we already have specialized devices for many tasks — dishwashers for dishes, Roombas for floors, and so on.

Robots Replace Function-Specific Devices, Not the Other Way Around

These comments echo the early skepticism about the internet, which reminds me of David Letterman’s 1995 interview with Bill Gates. Letterman questioned why the internet was needed, suggesting that radio and magazines could serve the same purpose, missing the transformative potential the internet would bring. Like the internet, robots may seem like redundant solutions today, but they could offer profound, emergent benefits that we can’t yet fully envision.

The appeal of humanoid robots lies in their ability to act as versatile, general-purpose solutions. Instead of requiring numerous function-specific devices, which can clutter and complicate life, robots have the potential to consolidate these tasks into a single, adaptable platform. Traditional consumers, though mostly unaware, hate the “brain lag” of juggling multiple devices and learning new interfaces. Consumers are far more likely to support a single, versatile solution than a constantly expanding set of specialized devices. This is why we have the internet instead of radios, magazines, the TV, and so on. This is also why robots will become integral to creating a more streamlined daily lifestyle. Furthermore, as robots become more integrated into our lives, they may unlock unforeseen benefits and efficiencies that can only emerge through widespread adoption and integration — just as the internet did.

Overcoming the Skepticism of Robots in the Home

Household robotics often face skepticism, with critics claiming they’re too complex or costly, better suited for industrial settings than everyday life. This mirrors early reactions to home computers, even among industry leaders. Ken Olsen, founder of Digital Equipment Corporation (DEC), famously stated, There is no reason for any individual to have a computer in their home. DEC was well-positioned to lead in personal computing but missed the opportunity due to this shortsighted view, allowing competitors like IBM to fill the rapidly growing niche with the now-dominant PC architecture.

Much like early computers, consumer robots may seem excessive now, but dismissing them as impractical or “only for the rich” overlooks the natural progression of technology. Just as computer prices eventually dropped to become accessible to the average household, advancements in robotics and decreased production costs will likely make robots viable companions for everyday tasks. As innovation drives accessibility, consumer robotics may evolve to provide convenience and functionality in the home, transforming daily life as personal computers once did.

Why Legs Instead of Wheels?

The question of why robots should have legs instead of wheels often centers on efficiency and practicality — wheels are simpler and excel on smooth terrain. However, by the time edge device inference and robotics research advance to the point of enabling robots to generalize across diverse, out-of-domain tasks, legged locomotion will likely be trivial in comparison.

Achieving generalization requires breakthroughs in handling complex, dynamic environments. For instance, interacting with soft or deformable items, understanding and manipulating unpredictable human movements, and operating dynamic interfaces like touchscreens all pose greater challenges than locomotion alone. As research in these areas progresses, so will advancements in legged mobility, allowing robots to adapt to diverse household and industrial scenarios, even those with stairs and uneven surfaces.

Futher, from a business perspective, the momentum around legged robots is building globally. Following Elon Musk’s Optimus announcement, numerous Chinese companies launched humanoid robotics projects focusing on legged designs. With China’s manufacturing strength driving down costs, legged robots are likely to become both affordable and accessible, creating a potential market advantage. To keep pace, U.S. robotics initiatives should prioritize legged locomotion as part of a broader strategy to build adaptable robots that can perform complex, real-world tasks effectively.

The Current State

Now, let’s address the second comment: research in humanoid robotics seems slow and non-generalizable.

Deeptech Industry Challenges

The progress in humanoid robotics may seem poor, with many robots primarily sectored off into controlled environments. However, this slow pace is inherent to tackling highly complex, unprecedented challenges. Consider Tesla’s achievements in autonomous driving or SpaceX’s work on reusable rockets — both of which have required years of iteration, research, and development to significantly reduce costs and increase reliability.

Notably, Tesla and SpaceX both entered markets with little to no serious competition, allowing them to set industry standards and push forward with ambitious, long-term projects. In Tesla’s case, while traditional automakers had EV research divisions, none were genuinely committed to mass electric vehicle production. Legacy car companies were hesitant to invest heavily in EV infrastructure, creating an environment in which Tesla could establish itself as a market leader with relatively minimal resistance. Similarly, when SpaceX entered the space industry, no other private companies in the U.S. were competing in orbital rocket production. SpaceX built the commercial spaceflight market from the ground up, reshaping the industry and setting new standards for reusable, cost-effective rockets. Both companies highlighted the immense value of tackling complex challenges within emerging fields and demonstrated the power of a singular vision and concentrated talent pool.

In contrast, the humanoid robotics industry is fragmented, with many companies operating in closed-source environments. Unlike Tesla or SpaceX, which centralized talent and expertise to create groundbreaking advancements, the robotics field is split among numerous companies, each striving to make similar progress in isolation. This dispersed approach leads to significant redundancy, with multiple reimplementations of the same foundational systems instead of a shared standard that all can build upon. Without a central leader to unify efforts, the field lacks the critical mass of talent and coordination necessary to drive rapid, cohesive advancements.

The Special Case of Figure

Figure’s media presence is an unusual phenomenon within the robotics industry and, in many ways, detrimental to the field’s progress. As a closed-source company, Figure has locked away its research and developments from the broader robotics community, limiting opportunities for shared advancement. Yet, despite this, Figure has captured the attention of the general public and investors alike, often being portrayed as the pinnacle of robotics companies. This perception is not only misleading but potentially harmful to genuine progress in the industry, as it reinforces the idea that Figure’s approach represents the highest standard in robotics. The reality, however, is quite different, and the best outcome might be for Figure to dissolve without creating a damaging ripple effect across the sector.

Figure’s CEO, Brett Adcock, raises additional concerns regarding the company’s leadership. Adcock has a history that suggests he may be more interested in capitalizing on industry hype than advancing robotics for its own sake. His previous ventures, such as Archer Aviation, reflect a pattern of using hype-driven markets to attract funding, with questionable moves like selling secondaries early on, which casts doubt on his commitment to long-term success. Rather than building companies as lifelong missions, Adcock’s public persona and business strategy imply a focus on establishing multiple companies for accolades and quick capital gains, treating each as a short-term success to be highlighted on a resume rather than as an enduring passion project.

A company without genuine leadership commitment to its mission is unlikely to succeed in such a challenging field. Robotics requires vision, resilience, and deep, focused dedication — qualities that are difficult to uphold without authentic belief in the company’s purpose. Figure’s current trajectory under Adcock risks misguiding investors and stalling real progress by occupying the market spotlight with promises that may not be backed by sustainable intent. For the robotics industry to thrive, companies must be led by individuals who truly understand and are driven by the unique challenges and long-term goals of this field, rather than those who see robotics as simply the next wave to ride.

So What’s the Point?

Humanoid robots represent the next evolution of AI itself, not just an adjunct to it. For AI to truly understand and engage with our human-centered world, it must inhabit a form capable of experiencing it similarly to us. Imagine a child trying to learn about human interactions and environments while constrained to a non-humanoid form — observing but never fully participating in the human experience. A humanoid form enables robots to directly learn and adapt through observation and interaction of other humans within our world’s dynamic contexts, making them not just tools but necessary extensions of advanced AI.

For the Investors

Investment in robotics should prioritize teams with proven experience in tackling deep-tech challenges and a demonstrated, long-term commitment to the field. Humanoid robotics demands resilience, specialized expertise, and vision — qualities best found in founders who have consistently tackled and solved complex problems in related domains. A team with a track record of navigating deep-tech hurdles is better positioned to address the multifaceted challenges inherent to robotics, from technical innovation to scalable manufacturing.

For the industry to reach its full potential, investment must either centralize into a select group of capable leaders or support open-source efforts that contribute to a shared knowledge pool. Such focused investment fosters rapid development, avoids unnecessary reinvention, and ensures that breakthroughs benefit the field as a whole.

For the Technical

Work on humanoid robots! There are many interesting, hard problems to solve. Even if you contribute just a little, working on hard problems and perfecting your solutions truly teaches you how to think different. If you move away from robotics, you’ll keep your new level of problem solving with you. And no matter what, you’ll be a better engineer.