
Humanoid vs purpose-built: What wins in robotics?
The Robotics Revolution: Why Purpose-Built Machines Lead the Way (For Now)
(This article was generated with AI and it’s based on a AI-generated transcription of a real talk on stage. While we strive for accuracy, we encourage readers to verify important information.)
At Web Summit Vancouver 2026, Daisy Xiong moderated a panel on humanoid versus purpose-built robots, exploring their future and current dominance. Kacper Nowicki, Co-founder & CEO of NoMagic.AI, defined humanoids as home-compatible, while industrial purpose-built robots, like single arms, have long succeeded in manufacturing and now automate logistics tasks.
James Wells, CEO of Sanctuary AI, stated humanoids lack critical performance for widespread usefulness, estimating 3-7 years for viability. Designing safe, high-degree-of-freedom robots with ROI is challenging, though physical AI advances rapidly, allowing application to off-the-shelf robots today.
Mr. Nowicki emphasized purpose-built robots’ immediate availability and 10+ year reliability. NoMagic.AI’s robots use tool changers to handle diverse warehouse items, from tiny to heavy, including delicate goods. This reflects significant progress in perception, decision-making, speed, and reliability for varied inventory, simplifying warehouse processes for customers.
Mr. Wells noted demand for purpose-built robots in dynamic, unstructured tasks, with manufacturers keen on automation. However, current adoption is small, representing only 0.5-1% of the $60 trillion global labor market. Industrial robots sell in hundreds of thousands annually, humanoids in tens of thousands, a tiny fraction of their potential.
The key challenge is achieving both extreme reliability (e.g., 99.9% success) and broad task generalization. Mr. Wells affirmed that generality, especially in manipulation, is the “holy grail.” Sanctuary AI’s Phoenix humanoid prioritizes bi-manual manipulation and dexterity for complex tasks, rather than just mobility, focusing on advanced robotic hands.
Both agreed zero-shot learning for robots remains a distant future. Factories, not homes, are the initial market for humanoids due to safety, existing infrastructure, and higher economic value. A logistics robot’s annual value ($150,000) far exceeds typical home robot budgets. People often overestimate robot capabilities from curated videos but underestimate genuine, rapidly advancing capabilities.
Mr. Nowicki added that true success is when robots become an unremarkable, reliable part of daily operations, like a “dusty robot” in a warehouse. NoMagic.AI’s zero churn and long-term deployments, with some robots operating for almost seven years, demonstrate this customer “addiction” to effective automation.

