+31 20 705 1010
EnglishNederlandsFrançaisDeutschEspañol
Models

World Labs launches Atlas to move AI from language to spatial world models

World Labs has introduced Atlas, a spatial intelligence model designed to move beyond text based processing. Unlike traditional large language models, this system focuses on creating world models that comprehend and interact with 3D physical spaces. The technology provides AI with a foundational understanding of depth, physics, and spatial relationships.

Why it matters

For enterprise teams, this shift represents a move from generative content to functional spatial reasoning. It provides the necessary framework for developing autonomous systems and robotics that require precise physical interaction rather than just linguistic simulation.

Key points

  • Atlas transitions AI from predicting text to predicting spatial dynamics
  • The model creates internal 3D representations from 2D visual inputs
  • Spatial intelligence enables more reliable automation in physical and simulated environments

Related solutions

AI Vision AI Agents & Automation Machine Learning
All AI news

More AI news

Models

Modern CISOs shift focus to enable strategic risk and accelerate AI adoption

Security leadership is moving away from a traditional gatekeeper role to become an active enabler of business risk and innovation. This transition focuses on guiding the safe integration of artificial intelligence while fostering a culture of responsible experimentation across the enterprise.

Models

Tech coalition urges US government to protect open-weight AI models

A coalition of major technology firms has petitioned the US government to safeguard the development of open-weight artificial intelligence models. The group argues that downloadable weights are essential for maintaining transparency and competition, ensuring that innovation is not restricted to a handful of proprietary providers.

Models

Lumia Lab scales latent space model to 8.9 billion parameters

LUMIA Lab has released the technical report for NCP-ArchPreview, an 8.9B parameter model trained on 5.73 trillion tokens. The model employs Next Concept Prediction to jointly predict tokens and latent representations, achieving performance parity with OLMo-3-7B while requiring significantly less data.

Ready to build something
extraordinary?

15 minutes. No pitch deck. Just a conversation about what AI can do for your team.

Talk directly with our AI specialists

15 min, no strings
No sales pressure
Prototype in 7 days