What happened
The 2026 AI100 Study Panel released its triennial report, 'AI Goes Mainstream: The Scaling Era and the Rise of .' The report highlights embodied AI and physical intelligence as a key open grand challenge, noting significant progress in vision-language-action models and simulation-to-reality transfer. It also reports that China leads in industrial robot deployment and produces approximately 80% of the world's humanoid robots, while emphasizing that a substantial gap remains between constrained demonstrations and broad real-world deployment due to the complexity of long-horizon planning and dexterous manipulation.
The One Hundred Year Study on Artificial Intelligence (AI100) released its 2026 report, titled 'AI Goes Mainstream: The Scaling Era and the Rise of .' This is the third edition of the longitudinal study, which assesses AI's development and societal impact every five years. The report is produced by an international committee of experts who address 12 standing questions regarding AI's technical, ethical, economic, and societal effects, comparing current developments with assessments from 2021.
A central focus of the 2026 report is the identification of six open grand challenges in AI. Among these, the panel highlighted physical AI and embodied intelligence as particularly inspiring and scientifically important. The report notes that over the past five years, there has been significant integration of natural language processing into robotic systems, the emergence of vision-language-action models for training and control, and advances in simulation technology that enable models trained in virtual environments to be deployed on physical hardware.
The report provides specific data on the state of industrial robotics, noting that China is the leader in this sector, deploying more robots than the rest of the world combined. It further states that China produces approximately 80% of the world's humanoid robots. While this sector currently accounts for only 2% of the market, the report characterizes it as growing rapidly.
Despite these advancements, the report emphasizes a significant gap between the demonstration of capabilities and actual deployment. Many current success stories are limited to highly constrained environments, such as warehouses, or involve very specific products like floor cleaners. The panel identifies the sheer complexity of the real world as the key challenge, citing specific scientific problems that need to be overcome, including long-horizon planning under uncertainty, achieving dexterous manipulation of diverse objects, and establishing formal performance guarantees.
The report also highlights a division of opinion among practitioners regarding the path forward. While some believe current statistical learning paradigms are sufficient to address these open problems, many others argue that fundamentally new ideas will be required to bring robot deployment into genuine real-world scenarios. The panel suggests that the next five years will be critical in observing how this debate plays out and whether new scientific approaches emerge.
Why it matters
The AI100 report serves as a high-level, expert-curated assessment of the field's trajectory, providing a structured framework for understanding where AI development is heading. By identifying embodied AI as a grand challenge, the report signals that the next phase of AI advancement will likely focus on bridging the gap between digital models and physical interaction. This has practical implications for robotics, manufacturing, and automation sectors, as it highlights the specific scientific hurdles—such as uncertainty in planning and dexterity—that must be overcome for widespread adoption. The data on China's dominance in hardware production further underscores the geopolitical and industrial dimensions of AI development, suggesting that future competition will be as much about physical infrastructure and deployment as it is about algorithmic breakthroughs.
The AI100 report provides a authoritative, expert-driven perspective on the direction of AI development, moving beyond immediate product releases to address long-term scientific and societal challenges. By designating embodied AI as a grand challenge, the report signals that the field's focus is shifting from purely digital generative models to systems that can interact with and manipulate the physical world. This shift has significant implications for industries reliant on automation, such as manufacturing, logistics, and healthcare, where the ability of AI to operate in unstructured environments is a critical bottleneck.
The report's findings on China's dominance in industrial robot deployment and humanoid robot production highlight the geopolitical and industrial dimensions of AI competition. With China producing 80% of the world's humanoid robots and deploying more industrial robots than the rest of the world combined, the report underscores that leadership in AI is not solely about algorithmic innovation but also about hardware manufacturing and deployment scale. This has practical implications for global supply chains and international policy, as nations may need to consider their own industrial capabilities in the context of AI development.
The identification of specific scientific hurdles, such as long-horizon planning under uncertainty and dexterous manipulation, provides a clear roadmap for researchers and engineers. These challenges are not merely technical but also fundamental to the reliability and safety of AI systems in real-world applications. The report's emphasis on the need for formal performance guarantees suggests that the industry is moving toward a more rigorous, safety-focused approach to AI development, which is essential for gaining public trust and regulatory approval.
The debate over whether current statistical learning paradigms are sufficient or if new ideas are needed reflects a broader uncertainty in the field. This uncertainty is important for investors, policymakers, and researchers, as it suggests that the path to general-purpose embodied AI may be longer and more complex than previously thought. The report's call for fundamentally new ideas indicates that incremental improvements may not be enough to overcome the challenges of real-world deployment, potentially leading to a shift in research priorities and funding allocations.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
What to watch next
Researchers and industry leaders should monitor developments in vision-language-action models and simulation technologies that aim to close the gap between virtual training and physical deployment. Additionally, the evolution of China's humanoid robot market, which currently represents a small but rapidly growing share of the industry, will be a key indicator of how quickly embodied AI moves from constrained environments to general-purpose applications. The report's emphasis on the need for potentially new scientific paradigms suggests that future breakthroughs may come from fundamental shifts in how AI systems learn and interact with the physical world.
Researchers and industry leaders should closely monitor developments in vision-language-action models and simulation technologies, as these are identified in the report as key enablers for bridging the gap between virtual training and physical deployment. Advances in these areas could lead to more robust and adaptable robotic systems that can operate in a wider range of environments.
The growth of China's humanoid robot market will be a key indicator of how quickly embodied AI moves from constrained environments to general-purpose applications. Given that this sector currently represents only 2% of the market but is growing rapidly, any significant expansion could have major implications for global robotics and automation industries.
The report's emphasis on the need for potentially new scientific paradigms suggests that future breakthroughs may come from fundamental shifts in how AI systems learn and interact with the physical world. Researchers should watch for new theoretical frameworks or algorithmic approaches that address the challenges of long-horizon planning and dexterous manipulation.
Policymakers and industry stakeholders should monitor the development of formal performance guarantees for AI systems, as this is identified as a key scientific problem that needs to be overcome. The establishment of such guarantees will be essential for ensuring the safety and reliability of AI systems in real-world applications, and could influence regulatory frameworks and industry standards.