The Partnership combines world-leading reality-grade and physically accurate 3D Digital Twin infrastructure with Protege’s real-world data platform for AI model development.

San Francisco, CA and New York, NY — ALLSIDES and Protege today announced a strategic partnership to advance the 3D data foundation powering the next generation of Physical AI, world models, robotics, and multimodal foundation models.

As AI moves beyond text, images and video into understanding and interacting with the physical world, access to high-quality, physically accurate, and scalable real-world 3D data has become one of the industry's most important challenges. Together, ALLSIDES and Protege are addressing this challenge by combining complementary capabilities at the intersection of 3D Digital Twin capture, data infrastructure, and AI development. The broader mission: to make every object and product in the world available to AI Labs building foundational models.

ALLSIDES has established itself as the global leader in reality-grade 3D Digital Twin infrastructure. The company's proprietary and automated 3D scanning technology enables the creation of highly accurate 3D digital twins of real-world objects at unprecedented quality and scale. Capturing not only visual appearance, but also true-to-life geometry, color, material properties, and physical characteristics like mass, colliders and friction of objects, ALLSIDES produces datasets that faithfully represent the physical world with a level of fidelity required for the next generation of AI systems.

Trusted by leading global brands and technology companies including Meta, Amazon, Nike, adidas, Zara, and Zalando, and working closely with partners such as NVIDIA, ALLSIDES is helping define the data layer for Physical AI. 

The company has also pioneered foundational work with Meta on the Digital Twin Catalog initiative, contributing to the creation of the world’s highest-quality datasets to date which is used by hundreds of researchers today in the fields of physical AI and robotics, computer vision and machine learning.

Today, ALLSIDES continues to build and scale this capability in-house, delivering reality-grade 3D data at enterprise scale, building a 3D Digital Twin dataset and infrastructure of millions of objects and products.

Protege gives leading AI model builders the right data that is purpose-built for AI model training and development. The company works directly with leading AI labs, robotics companies, and frontier model developers to source, refine, curate, and operationalize high-value datasets for training and evaluation, drastically shortening the time from raw data to purpose-built datasets for training and evaluation. Protege bridges the gap between proprietary data owners and organizations building the world’s most advanced AI models and applications.

Through this partnership, Protege expands its proprietary data platform to include the world's highest-fidelity 3D Digital Twin data representing real-world objects, unlocking the co-production of reality-grade and physically accurate 3D Digital Twins for leading frontier AI labs and robotics companies. 

–

Why this partnership matters

📦 Access to reality-grade 3D Digital Twin data representing real-world objects and products with unprecedented physical fidelity

🤖 Improved training and evaluation data for Physical AI, world models, robotics, and multimodal foundation models

🔬 Accelerated development of embodied AI systems capable of understanding and interacting with the physical world

"Over the past years we’ve been working with lots of researchers at AI Labs, robotics companies and industrial leaders and it is very clear that 3D AI, or physical AI, is much harder to develop then any text or image related AI model," said Franz Tschimben, CEO of ALLSIDES. "By partnering with Protege, we are combining best-in-class 3D Digital Twin technology with one of the strongest AI infrastructure players. Together, we provide unprecedented quality, scale, distribution and reach to empower and enable AI researchers across the globe to develop the next wave of groundbreaking AI models."

"Physical AI has a data problem that better models alone cannot solve," said Grant Murphy-Herndon, General Manager of Spatial & Physical Intelligence at Protege. “Capturing the real world in 3D – geometry, materials, mass, friction – is hard, and even once you've captured it, raw data is rarely ready to train on. ALLSIDES produces some of the most physically accurate digital twins anywhere, and Protege turns them into training-ready datasets so teams building world models and robots can build, not wrangle data."

The partnership reflects a shared belief that increasingly powerful models alone are not enough. The future of AI will depend on access to scalable, representative, and physically accurate data that faithfully captures the complexity of the real world.

‍