Humanoids: from Spectacle to Scale | Bloomberg Tech: Asia 5/29/2026

Key Concepts

• Humanoid Robotics: Robots designed with human-like physical forms to operate in environments built for humans.
• Embodied Intelligence: The integration of AI (like Large Language Models) into physical hardware, allowing robots to perceive, reason, and act in the real world.
• Foundation Models (VLA): Vision-Language-Action models that enable robots to process visual data and natural language to execute complex behaviors.
• Dexterity: The ability of a robot to perform fine-motor tasks (e.g., folding origami, packing lunchboxes) using multi-fingered hands.
• Functional/Semantic Safety: The multi-layered approach to ensuring robots are stable, avoid collisions, and possess "common sense" to navigate unstructured environments.
• Teleoperation vs. Autonomy: The transition from remote-controlled robots to fully autonomous systems capable of independent decision-making.

---

1. Market Outlook and Economic Impact

The humanoid robotics industry is transitioning from "proof of concept" to real-world deployment.

• Market Valuation: Currently estimated at $2–3 billion, forecasts suggest growth to $40 billion by 2035, with a potential long-term impact of $5 trillion by 2050 when accounting for supply chains and services.
• Investment Trends: Venture funding has surged from $700 million in 2018 to over $4 billion in 2023, with startups like Figure AI and Unitree achieving billion-dollar valuations.
• Deployment Projections: While only 13,000 humanoids were shipped in 2025, projections estimate 12 million units deployed by 2035.

2. Technological Evolution and Frameworks

The industry is moving away from "predefined, memorized sequences" toward AI-driven adaptability.

• The Shift to Data-Driven Robotics: The introduction of Transformers and Foundation Models has shifted the field from hard-coded programming to learning-based systems.
• Reinforcement Learning (RL): Used for whole-body control, allowing robots to perform complex acrobatics and adapt to unstructured spaces.
• The "Edge" of Intelligence: The primary challenge is enabling robots to understand the nuance of human environments, which are constantly changing and unstructured.
• Hardware Trajectory: Experts emphasize that the "unblocking" of the industry requires combining accelerated hardware development with advanced AI reasoning.

3. Key Industry Players and Perspectives

• Google DeepMind: Focuses on "Embodied Intelligence" by integrating the Gemini model into robotics. Their goal is to move beyond simple tasks to highly dexterous manipulation.
• Honda R&D: A pioneer in bipedal locomotion (e.g., ASIMO), Honda has shifted its focus toward "Avatar Robotics"—a vision of 4D mobility where robots can be teleoperated or delegated to perform tasks across time and distance. They emphasize the importance of multi-fingered hands for human-like interaction.
• Global Competition:
  • China: Leads in volume and cost-reduction.
  • USA: Leads in AI and software intelligence.
  • Japan: Maintains a competitive edge in high-quality, precise hardware manufacturing.

4. Challenges to Widespread Adoption

• Labor Shortages: The primary driver for adoption is the global need to fill labor gaps in manufacturing and services.
• Cultural Acceptance: Adoption rates vary by region; a Mitsubishi Research survey indicated 56% of Chinese respondents are open to humanoids in daily life, with lower acceptance rates in Japan and the US.
• Safety and Security: There is a lack of standardized safety protocols. Concerns include physical safety (stability/collision avoidance), "semantic" safety (common sense), and cybersecurity (preventing remote hacking).
• Scalability: The industry must move from "gimmicky" demonstrations to sustainable business models that generate profit.

5. Notable Quotes

• Carolina Parada (Google DeepMind): "Embodied intelligence is the next frontier on AI... it brings all of Gemini's world understanding, multimodality in order to enable robots to understand their environment, to reason, and to be able to take action to the level of precision of a human expert."
• Takahide Yoshi (Honda R&D): "The Chinese company tech companies have strength in cost reduction... the US company has strength in AI... Japanese companies have strength in making precise or high-quality hardware."
• Ian Maya (Bloomberg Intelligence): "We are at the point where companies are starting to realize which type of humanoid will be able to help them make money and be able to sustain that as a business."

Synthesis/Conclusion

The humanoid robotics sector is at a critical inflection point. While the "hype" is currently driven by impressive demonstrations, the path to a $5 trillion industry relies on solving the "dexterity gap" and achieving true autonomous reasoning. The future of the industry will likely mirror the smartphone market, with a tiered ecosystem ranging from low-cost, simple-task robots to premium, highly sophisticated machines capable of complex industrial and household labor. Success will depend on the convergence of high-quality hardware, robust AI foundation models, and the establishment of global safety and security standards.