BEIJING, May 7, 2026 /PRNewswire/ -- Spirit AI (千寻智能) and Bosch China have formed a strategic alliance to industrialize the "Universal Brain" for robots. This partnership fuses Spirit AI's leading VLA models with Bosch's deep industrial ecosystem to bridge the gap between frontier embodied AI and large-scale industrial deployment.

Key Highlights of the Partnership:

• The Data Loop: Establishing a "Real-World Data — Embodied AI — Real-World Scenarios" pipeline within Bosch's China-based factories and logistics centers.
• Hardware Synergy: Integration of Bosch's mission-critical sensors and actuators to accelerate Spirit AI's engineering validation and mass production.
• Embodied AI Model Iteration: Benchmarking Spirit v1.5 against global standards to ensure peak performance in complex, unstructured industrial tasks.

"This alliance marks our transition from technical excellence to industry leadership. Bosch's industrial footprint across real-world scenarios, global supply chain, and channel resources will scale the reach of our 'Universal Brain' architecture." — Han Fengtao (韩峰涛), Co-founder & CEO of Spirit AI

"Spirit AI has demonstrated leading technical prowess in the field of embodied AI models," stated Liu Min, VP of Strategic Development at Bosch China and Head of the Bosch China Robotics Center. "Our collaboration will establish a new ecosystem paradigm for the robotics industry."

About Spirit AI

Founded in 2024, Spirit AI is dedicated to developing "universal brains" for robots to create the next generation of intelligent workforce. By deploying general-purpose embodied models, the company provides robots with the robust generalization and physical precision required for the real world. Spirit AI is moving beyond the lab to integrate versatile robotic agents into the modern workforce, accelerating the arrival of real-world embodied AI.

Contact: pr@spirit-ai.com 

BEIJING, Feb. 25, 2026 /PRNewswire/ -- Spirit AI has raised $280 million USD to scale the deployment of general-purpose embodied models. The funding arrives as the industry pivots toward Scaling Law-driven VLA architectures—a trajectory supported by a diverse group of global financial and strategic investors.

This Beijing-based company is building a universal robotic brain by scaling with diverse human video and wearable sensor data. This path aligns Spirit AI with global peers like Google DeepMind and Physical Intelligence (Pi) in leveraging massive datasets for physical reasoning. The vision is powered by a core team from UC Berkeley, Tsinghua, and Peking University — averaging under age 30—who bridge frontier theory in multimodal LLMs and robot learning with industrial-scale deployment.

The "Dirty Data" Strategy: Scaling Beyond Curation

While many in the field have hit performance ceilings by over-curating "clean" datasets, Spirit AI is prioritizing real-world complexity. "Dirty data is the key to scaling VLA models," says Yang Gao, Co-founder & Chief Scientist of Spirit AI.

Dr. Gao currently serves as an Assistant Professor at Tsinghua University and holds a PhD from UC Berkeley. A prominent figure in robot learning, he has spearheaded a range of influential research while bridging academia and industry. His notable contributions include EfficientZero, scaling laws for imitation learning, and pioneering frameworks such as ViLa and CoPa.

The company argues that diverse, unstructured, and non-pre-scripted interaction is the essential catalyst for building models with true common sense.

-          Data Velocity: Spirit AI has amassed over 200,000 hours of interaction data, with a roadmap to exceed 1 million hours by the end of 2026.

-          Cost Disruption: Using proprietary wearable collection devices, Spirit AI has reduced data acquisition costs by 90% compared to traditional teleoperation.

-          Benchmark Performance: In January 2026, Spirit v1.5 topped the RoboChallenge global leaderboard, demonstrating state-of-the-art generalization that rivals the world's leading embodied AI models.

Industrial Validation: The CATL Benchmark

Spirit AI has applied VLA models to the production lines of CATL, the world's largest battery manufacturer.

On the floor, Spirit AI-powered agents handle flexible wire harnesses—a long-standing hurdle due to material unpredictability. Achieving a 99%+ success rate, these agents match the precision and cycle times of skilled human workers in complex manufacturing.

About Spirit AI

Spirit AI builds the "Universal Brain" for the next generation of robotics. By deploying general-purpose embodied models that bridge simulation and reality, the company provides robots with the robust generalization and physical precision required for the real world. Spirit AI is moving beyond the lab to integrate versatile robotic agents into the modern workforce, accelerating the arrival of real-world embodied AI.

Media Contact: pr@spirit-ai.com

RoboChallenge Leaderboard: https://robochallenge.cn/home

Open Source: 

Code: https://github.com/Spirit-AI-Team/spirit-v1.5
Model: https://huggingface.co/Spirit-AI-robotics/Spirit-v1.5
Blog：https://www.spirit-ai.com/en/blog/spirit-v1-5

Spirit v1.5 was evaluated on RoboChallenge Table30. RoboChallenge is a standardized real-robot evaluation benchmark jointly initiated by organizations including Dexmal and Hugging Face, with the goal of assessing embodied AI systems under realistic execution conditions.

The tasks span everyday skills such as object insertion, food preparation, and multi-step tool use, and are evaluated across multiple robotic configurations, including single-arm and dual-arm systems with varying perception setups. The benchmark is designed to stress a model's ability in 3D localization, occlusion handling, temporal reasoning, long-horizon execution, and cross-robot generalization.

A Unified Vision-Language-Action Model for Real-World Execution

Spirit v1.5 is built on a unified Vision-Language-Action (VLA) architecture that integrates visual perception, language understanding, and action generation into a single end-to-end decision process. Unlike modular pipelines that separate perception, planning, and control, this unified approach reduces information loss and enables more consistent behavior across complex, multi-stage tasks.

A key technical focus of Spirit v1.5 is its data collection paradigm. Rather than relying on highly curated, scripted demonstrations, Spirit v1.5 is largely trained on open-ended, goal-driven diverse data, where operators pursue high-level objectives without predefined action scripts. This paradigm allows training data to naturally capture a continuous flow of skills, including task transitions, recovery behaviors, and interactions across varied objects and environments.

By learning from this unstructured and diverse experience, the model develops more transferable and generalizable policies, which later translate into stable performance on complex, multi-stage robotic tasks evaluated in real-world benchmarks.

Training on Diverse, Unscripted Real-World Data

In this data collection paradigm, operators are given high-level goals rather than scripted action sequences, allowing tasks to unfold naturally and organically. As a result, a single data session may contain a continuous stream of diverse atomic skills—such as grasping, inserting, twisting, opening containers, and coordinated bimanual actions—closely resembling real human environments.

This diversity enables the model to learn not isolated behaviors, but how skills connect and transition, forming a more general and transferable policy.

Improved Generalization and Transfer Efficiency

Results from recent ablation studies reveal a notable correlation between pre-training data variety and transfer efficiency. According to the data, models exposed to diverse, unscripted content during pre-training require significantly less time to master novel tasks during fine-tuning than their counterparts trained on scripted demonstrations. This efficiency gain was observed while maintaining identical data budgets across both cohorts.

These results suggest that task diversity, rather than task purity, is a critical driver for scalable embodied AI. As the volume of diverse experience increases, Spirit v1.5 continues to show improved performance on new tasks, supporting its role as a general-purpose embodied foundation model.

Open-Source Release and Reproducibility

In a move toward industry transparency, Spirit AI has released the model weights and source code utilized for the RoboChallenge evaluation. The open-source availability of these assets allows the research community to independently verify benchmark results. Furthermore, it provides a foundational framework for developers to extend Spirit v1.5, potentially accelerating advancements in embodied intelligence and robotics research.

About Spirit AI

Website: https://www.spirit-ai.com/en/ 

Spirit AI is a leading frontier startup dedicated to building the "universal brain" for embodied AI. The company focuses on developing advanced embodied large models to create general-purpose robotic companions for every household. By bridging cutting-edge AI with physical interaction, Spirit AI is driving the global transition toward the era of intelligent robotics.