De-romanticising the robot: A reality check in the age of billion-dollar humanoid dreams

The robotics industry finds itself at a fascinating crossroads. While Figure AI recently secured more than $1 billion at a staggering $39 billion valuation—a 15x increase from its previous funding round—and other humanoid robotics companies chase similarly astronomical valuations, some industry practitioners advocate for a more grounded approach to automation.

Gokul NA and Nikhil Ramaswamy, the founders of CynLr, a Bengaluru-based robotics company that has deliberately chosen a different path, understand this disconnect between headline- and valuation-grabbing fervour and operational reality. Their philosophy, gleaned from years of industry experience, starts with the end-customer user’s problems and solves 90 per cent of the tasks that aren’t automated today.

This approach is rooted in solving real-world problems—on factory floors and warehouses, for example—and not simply the form factor.

The appeal of humanoid robots is undeniable. They capture imaginations, dominate headlines, and attract venture capital. The reasoning appears sound: since we've built our world for human proportions and capabilities, robots that mirror our form should seamlessly integrate into existing environments.

This logic, while compelling in theory, does not yet pass practical scrutiny. “There's a lot of noise around what's happening in robotics,” notes Ramaswamy. "There are humanoids, there's AI, there's a lot of promise. And sometimes the promise, while it's exciting, creates a lot of mistrust about whether this tech is being looked at fundamentally.”

Before establishing CynLr in 2019, they spent four years consulting in machine vision and robotics, working directly with manufacturers to solve automation challenges. This ground-level perspective revealed a crucial insight: the problem isn't the robot's appearance—it's the robot's capabilities.

CynLr's three-arm robotic system, conceived well before the current humanoid wave, emerged after analysing what manufacturing tasks actually require. “We imagined well before humanoid took into play, and then we started talking about generalisation of hardware and general-purpose robotics much before the humanoid came in,” explains Gokul.

This philosophy extends to their broader vision of “software-defined factories” — manufacturing environments that can be reconfigured through programming rather than physical restructuring. Such flexibility addresses real industry pain points: the inability to quickly adapt production lines to changing consumer demands or new product designs.

Consider the practical benefits: Traditional automation requires extensive customisation, consuming 15–36 months for design and validation. By the time implementation is complete, product specifications may have already changed. CynLr's approach enables robots to handle unfamiliar objects through real-time learning, eliminating the need for exhaustive pre-programming.

The current funding environment reveals a troubling pattern: massive investments flowing toward companies promising human-like robots while fundamental intelligence challenges remain unsolved. Ramaswamy advocates for “de-romanticising this word, robot,” arguing that the focus should be on utility creation rather than anthropomorphic aesthetics.

“From the world of robotics, the utility really is to relieve human beings of labour that involves manipulating the real world and objects in the real world,” he explains. The form factor should emerge from function, not the reverse. A dishwasher doesn't need to look human to clean dishes effectively — it needs to perform the task reliably and efficiently.

This perspective gains credence when examining actual deployment challenges. Even advanced humanoid systems struggle with basic object manipulation tasks that humans perform instinctively. The gap isn't in mechanical design but in perception and intelligence — the ability to understand, adapt, and learn from complex, unstructured environments.

The venture capital enthusiasm for humanoid robotics reflects broader market dynamics rather than manufacturing requirements. As Gokul notes, it’s not uncommon for VC funding to follow templates, but deep tech companies require patient capital and foundational thinking rather than trend-following.

This mismatch creates distorted incentives. Companies may prioritise features that generate media attention and investor interest rather than solve operational problems. The result is a disconnect between what captures headlines and what addresses real industrial needs.

Meanwhile, genuine automation challenges persist. Despite decades of robotics development, even sophisticated manufacturing facilities operate with minimal or sub-optimal levels of automation. A typical automaker's plant has less than 10 per cent of its tasks automated, leaving 90 per cent dependent on manual labour. This statistic alone suggests that the industry should focus its attention on practical solutions instead of humanoid ambitions.

CynLr's decision to establish a research centre in Switzerland reflects their commitment to foundational technology development rather than market hype. “The whole industry is still grappling to see what they are supposed to do,” notes Gokul, referencing recent geopolitical uncertainties and supply chain disruptions.

CynLr’s approach mirrors historical technology evolution. Prior to the large-scale development and adoption of operating systems, hardware standards had to be stabilised. Before widespread software adoption became possible, fundamental computing infrastructure required development. Similarly, universal robotic manipulation may require foundational breakthroughs in perception and intelligence rather than anthropomorphic form factors.

This isn't an argument against ambitious robotics research or significant investment in automation technologies. Rather, it's a call for measured expectations and practical focus. The robotics industry would benefit from more companies pursuing CynLr's philosophy: understanding customer problems deeply, building foundational technologies systematically, and resisting the temptation to chase trends rather than solve real challenges.