They don’t dance or leap like sci-fi androids, but these utilitarian robots are proving their worth and attracting major investment by doing one thing well: handling specific tasks efficiently and at low cost.
Unlike general-purpose humanoids designed for complex, adaptive jobs, these boxy, function-first machines are built for clear purposes such as hauling goods, collecting trash, or inspecting equipment. Their simplicity and reliability are fueling a shift in robotics investment from flashy innovation to practical application.
As industries ranging from retail and defense to waste management ramp up automation, investors and analysts say specialized robots offer a more direct path to profitability. Meanwhile, humanoid robot developers continue to wrestle with high costs, limited data, and challenges operating in unpredictable environments.
According to PitchBook, global robotics startups raised $2.26 billion in the first quarter of 2025, with over 70% of that capital going to companies focused on task-specific machines.
This surge in funding is part of a broader global competition in robotics, with China emerging as a strong contender due to significant government support. Advances in chip technology are also accelerating progress, enabling robots to run complex AI models directly on-device, without needing cloud-based processing.
With Nvidia’s Orin NX, we can run many more AI models on the edge than before,”said Saurabh Chandra, CEO of India-based Ati Motors. The company’s flagship Sherpa Tug robot, which can pull over 1,000 kg in factories and industrial settings, has clocked more than 500,000 kilometers of autonomous operation and is deployed at companies like Hyundai, Forvia, and Bosch.
In healthcare, Diligent Robotics of Austin is gaining traction with Moxi, a robot designed for non-clinical hospital tasks such as delivering supplies and lab samples. “By solving a specific, high-need problem in healthcare, we’ve built a sustainable business,” said CEO Andrea Thomaz, noting that Moxi has reached product-level profitability.
Challenges for Humanoids
While interest in robotics soars, humanoid robots continue to struggle. Teaching robots to navigate unpredictable environments and develop real-world reasoning remains difficult. Unlike generative AI models trained on massive datasets, humanoid robots need physical-world interaction and task-specific training data.
Firms like Figure AI are working on this challenge, aiming to ship 100,000 humanoid robots in the next four years. Their robots rely on advanced AI to process real-time sensory input but are mostly confined to structured settings like car factories.
Cost remains another barrier. Components such as lidar and cameras push the cost of humanoids to $50,000–$200,000 per unit, compared to $5,000–$100,000 for specialized task-focused machines, according to a market analysis by Standard Bots.
True general-purpose robots haven’t been invented yet, said Marc Theermann, Chief Strategy Officer at Boston Dynamics. He warned that promises of commercially viable general-purpose robots are premature and likely to under-deliver.
Read also:
Boston Dynamics’ four-legged Spot robot, for instance, has found success in a narrow but profitable niche: hazardous industrial inspections. Meanwhile, its humanoid robot Atlas may go viral for flips and dances, but it’s not yet commercially viable.
Investors are taking note. Era Ventures has backed ViaBot, which automates trash collection in parking lots, while Parkway Venture Capital has split its bets between forklift automation startup Siera AI and Figure AI’s humanoid development.
We’ll see more robots built to do one task well and cost-effectively, said Raja Ghawi, Partner at Era Ventures. Eventually, that progress may lay the foundation for a fully functional humanoid robot.
