Global Objects Announces Strategic Collaboration with Microsoft to Build the Ground-Truth Foundation for Physical AI on Microsoft Azure

Global Objects, Inc. (GO), a spatial intelligence company digitizing the physical world, announced a multi-year strategic collaboration with Microsoft to build a retrieval-grounded generative AI world model on Microsoft Azure trained on real, multimodally scanned, licensed physical, 3D objects, generative AI world model rather than internet-scraped video. The program will scale to an initial seed corpus of 1.67 million unique licensed objects. As part of the collaboration, GO Scout, the first product in the Global Objects ecosystem, is now available in the Microsoft Marketplace, with additional products to follow.

Also Read: AiThority Interview with Matej Bukovinski, Chief Technology Officer at Nutrient

Strategic Microsoft Collaboration Will Build Retrieval-Grounded Generative AI World Model Trained on Real, Licensed, Physical 3D Objects

The biggest bottleneck in Physical AI today is not compute, it is data. The verified, physics-accurate ground truth that lets robots, autonomous systems, and industrial AI perform reliably in the real world is scarce and rarely licensed for commercial use. Global Objects is building the substrate that closes that gap, and the ecosystem of products that puts it to work.

“Today’s generative systems hallucinate physics from pixels. We are building the opposite: a world model trained on real objects measured, scanned inside and out, and licensed at the source,” said Jess Loren, Co-Founder and CEO of Global Objects. “Surface geometry is just the outer shell. To give AI a true understanding of an object, you need its mass, its articulation, how it deforms under force, what it sounds like when it falls. With Microsoft Azure as the foundation, we will make grounded spatial intelligence as accessible as a language model and the GO ecosystem is how we put it into customers’ hands.”

A Multimodal Substrate for Physical AI

Each object in the GO corpus is captured across the modalities that matter to a downstream AI system:

• Geometry and materials — High-resolution photogrammetric scans and PBR material maps.
• Internal structure — CT scan data revealing what is inside the object, not just its surface.
• Mass and physics — Measured weight, density, articulation, friction, and deformation behavior.
• Acoustic and thermal — How an object sounds when struck or moved, and how it conducts heat.

A public preview of the GO Physical Artifact Corpus (GO-PAC) is available at globalobjects.com.

What the Collaboration Delivers

• Multimodal corpus and generative world model on Azure — physically accurate, IP-clean, and auditable by construction.
• GO Scout in the Microsoft Marketplace — the first product in the Global Objects ecosystem, available now for enterprise procurement.
• Azure-native spatial intelligence services — 3D scene reconstruction, sensor-fused world models, and physics-validated training data on Azure.
• Joint go-to-market — co-developed reference solutions and sales motions across media and entertainment, robotics and physical AI, industrial digital twins, retail and consumer goods, and automotive.
• Responsible AI provenance — full data lineage from physical object capture through model output, aligned with Microsoft’s Responsible AI framework.

“Access to Global Objects physical-world data gives Microsoft Azure customers a decisive advantage with Physical AI. It connects AI to reality, so insights reflect how infrastructure, supply chains, or even movie sets actually behave, not just how models predict they might,” said Alistair Speirs, General Manager, Global Platforms at Microsoft. “Through our collaboration with Global Objects, Azure customers gain access to verifiable, high-quality datasets grounded in the real world, so they can put Physical AI to work with the security, scale, and trust needed to transform their industries.”

Also Read: ​​AI systems – Interoperable AI systems: Connecting models across platforms

[To share your insights with us, please write to psen@itechseries.com]