# OpenAI's Robotics Team Disbanded Amid Shift to General AI Models

> In a bold reorganization, OpenAI has dissolved its dedicated robotics division, channeling resources into broader AI initiatives. This pivotal decision marks a new chapter in the company's approach to robotics, leveraging general AI capabilities.

**Source**: fortune.com | **Published**: 2026-11-13 | **Type**: article

## Key Facts

- OpenAI's $10.9B quarterly losses highlight urgent need for cost-cutting and focus on core products.
- GPT-6 Astra's 74.8% success rate vs. 51.0% for specialized models indicates a shift in robotics strategy.
- Disbanding robotics team suggests OpenAI views robotics as a downstream application of general AI.
- Competitive vulnerability: robotics startups may struggle as general models challenge their foundational premise.
- Reallocation of researchers signals strategic pivot towards integrated AI solutions over specialized robotics.

## Summary

OpenAI has recently disbanded its dedicated robotics team, reallocating its researchers to broader projects within the company. This move signals a strategic shift in how OpenAI approaches robotics, emphasizing the integration of general-purpose AI models into robotic applications rather than maintaining a separate robotics division. The company maintains that robotics remains a core aspect of its long-term vision, suggesting this reorganization is not a retreat but a recalibration of focus.

The reorganization comes at a critical juncture for OpenAI, which faces mounting pressure to streamline operations and reduce costs. In the last quarter, the company reported significant operating losses of $10.9 billion. As OpenAI prepares for a potential IPO in early 2027, CEO Sam Altman has urged the company to concentrate on its core products, leading to the discontinuation of several projects, including the Sora video generation model. By merging the robotics team into its main research and applied divisions, OpenAI aims to enhance efficiency and leverage its advancements in AI technology.

Historically, OpenAI's foray into robotics has been fraught with challenges. The company initially disbanded its robotics team in 2020, citing insufficient data and a strategic pivot towards developing artificial general intelligence (AGI) through large language models. A brief revival occurred in 2024 with a partnership aimed at humanoid robotics, but that collaboration ended in February 2025, with Figure opting to develop its own AI solutions. OpenAI is now reportedly working on its humanoid robot, with a public demonstration slated for 2027.

The shift in strategy reflects a growing belief within the AI industry that general-purpose models can effectively manage robotic tasks, potentially eliminating the need for specialized robotics teams. OpenAI's recent experiments with its GPT-6 Astra model have shown promising results, with the model successfully controlling robotic functions in simulated environments. This approach challenges the traditional view that robotics requires tailored AI solutions, as many startups have raised substantial funding based on the premise that robots need dedicated intelligence.

Experts are divided on the implications of this shift. Guanya Shi from Carnegie Mellon University interprets OpenAI's reorganization as a recognition that robotics is no longer a distinct challenge but rather an application of general intelligence. Conversely, Russ Tedrake from MIT emphasizes that robotics companies possess a unique advantage through real-world deployment experience, which allows them to gather valuable data and adapt their robots to specific environments and tasks. The insights gained from these deployments could enhance the effectiveness of general-purpose models when applied in practical settings.

Despite the potential of general AI models, challenges remain. Large language models, while powerful, require constant internet connectivity and can be slow, which may necessitate the continued development of specialized models capable of operating independently on robotic hardware. The future of robotics may involve a hybrid approach, where general models are complemented by smaller, specialized systems that can function autonomously.

OpenAI's decision to integrate its robotics researchers into broader teams reflects a strategic pivot that could reshape the competitive landscape in robotics. As the capabilities of general-purpose models expand, the traditional barriers between AI and robotics may blur, leading to new opportunities and challenges. Companies in the robotics sector must adapt to this evolving paradigm, reconsidering their reliance on specialized models and exploring how they can leverage advancements in general AI to enhance their offerings. The implications of this shift could redefine the future of robotics, as firms that successfully integrate general intelligence into their operations may gain a significant competitive edge.

## Entities

- **Companies**: OpenAI, Figure, Nvidia, Generalist
- **Products**: ChatGPT, GPT-6 Astra, Sora
- **Technologies**: artificial general intelligence, large language models, embodied AI
- **People**: Sam Altman, Mark Chen, Greg Brockman, Wojciech Zaremba, Guanya Shi, Russ Tedrake, Pete Florence
- **Organizations**: Carnegie Mellon University, Walden Robotics

## Key Concepts

reorganization of OpenAI's robotics team, general-purpose AI models, embodied AI, robotics and AI integration, cost management in AI companies, deployment experience of robots, performance of general models in robotics, future of robotics startups

## Definitions

- **artificial general intelligence**: A type of AI that aims to replicate human-like cognitive abilities across a wide range of tasks.
- **embodied AI**: AI models that control physical devices, enabling them to act in the real world.
- **large language models**: AI models designed to understand and generate human language, often trained on vast datasets.
- **deployment experience**: The practical experience gained by robots operating in real-world environments, which helps them adapt to specific tasks.
- **general-purpose models**: AI models that can be applied to various tasks without needing specialized training for each specific application.

## Use Cases

- controlling robots with general-purpose AI models
- developing humanoid robots
- using AI for factory automation
- enhancing robot adaptability through deployment experience
- applying AI in driverless cars
- integrating AI in drones

## Frequently Asked Questions

**Why did OpenAI disband its robotics team?**

OpenAI disbanded its robotics team to reassign researchers to focus on general-purpose AI models, which they believe can effectively control robots without the need for specialized models.

**What is the significance of general-purpose AI models in robotics?**

General-purpose AI models can perform tasks traditionally requiring specialized training, potentially streamlining the development of robotics and reducing the need for separate robotics teams.

**What challenges do general-purpose models face in robotics?**

While general-purpose models show promise, they often require a constant internet connection and may be slower compared to specialized models that can operate offline.

**How does OpenAI's reorganization reflect industry trends?**

The reorganization aligns with a growing belief in the AI industry that general-purpose models can be effectively applied to robotics, challenging the need for dedicated robotics teams.

**What future developments can we expect from OpenAI in robotics?**

OpenAI is currently developing its own humanoid robot, with plans for a public demonstration in 2027, indicating ongoing commitment to robotics despite the team reorganization.

## Links

- [Read on Welcome.AI](https://welcome.ai/content/openais-robotics-team-disbanded-amid-shift-to-general-ai-models)
- [Original source](https://fortune.com/2026/11/13/openai-disbands-robotics-team-as-its-general-models-start-running-robots/)

---

Source: Welcome.AI | https://welcome.ai/content/openais-robotics-team-disbanded-amid-shift-to-general-ai-models