---
title: "Niantic Shifts Focus Towards Developing World Models for Simulations"
url: https://newscentral.site/niantic-shifts-focus-towards-developing-world-models-for-simulations/
language: en
publisher: "News Central Site"
section: "Technology"
published: 2026-11-08T10:00:00.000Z
updated: 2026-11-09T06:55:58.558Z
id: 72925055-4f34-4ad6-93b3-da63c8a0f597
source: "Wired https://www.wired.com/story/world-models-will-soon-be-inescapable-what-you-need-to-know/"
attribution: "Link to https://newscentral.site/niantic-shifts-focus-towards-developing-world-models-for-simulations/ and name News Central Site when you quote or summarize this story."
---

# Niantic Shifts Focus Towards Developing World Models for Simulations

Niantic has shifted its focus towards developing world models, a type of AI that reconstructs and simulates real-world environments using visual inputs.

Niantic, the company behind Pokémon Go, has shifted its focus towards developing world models, a type of AI that reconstructs and simulates real-world environments using visual inputs.

These models can generate digital environments that mimic the behavior of physical objects, allowing for simulations to be run in a virtual space. According to Brian McClendon, chief technology officer at Niantic Spatial, this means creating replicas of the world that are indistinguishable from reality.

The potential applications of world models are vast and varied, with several industries poised to benefit from their development. Home robotics and driverless cars, for example, could be significantly advanced in 2027 due to the use of world models.

World models differ from large language models in that they rely on visual inputs rather than text-based data. This allows them to generate three-dimensional environments that can be used as digital training grounds for robots and other autonomous systems.

The use of world models is not limited to humanoid robots, as driverless cars also benefit from simulated practice environments.

For autonomous vehicles, swift software reactions are crucial in high-pressure situations where human intervention is not possible.

Jamie Shotton, chief scientist at Wayve, a company that develops world models for cars, emphasizes the importance of a world model's ability to respond to actions taken in the physical environment. It must comprehend fundamental concepts such as physics and cause-and-effect relationships.

Wayve's generative world model draws from dashcam footage to recreate hazardous road conditions and other scenarios that are impractical or unsafe to test in real life.

Developers are also exploring how world models could enhance video games, given their reliance on physics and 3D rendering capabilities. Major players like Tencent are investing in this technology.

Large AI companies, such as Google, will likely continue to tout advancements in world-model development. Meanwhile, investors are pouring money into startups that focus on computer science innovations related to world models.

Yann LeCun, former chief AI scientist at Meta, recently secured over a billion dollars for his Advanced Machines Intelligence startup, which is dedicated to developing this technology.

Fei-Fei Li, a renowned Stanford computer scientist, launched World Labs in 2024. This company was later acquired by AMD for more than $8 billion this year.

The influence of world models is spreading beyond the tech industry, affecting workers in various sectors as companies seek to improve their simulations. One way AI labs gather data for these models is by recruiting individuals to record their daily activities.

People are being hired to perform mundane tasks and document them on camera, with some even selling footage through apps. The work is often low-paying, but it contributes to the development of more sophisticated world models. As a result, gig workers may find themselves in higher demand as companies seek to collect more data.

The proliferation of world models has led to an unsettling feeling among those who have contributed to their creation, they can't shake off the image of humanoid robots trapped inside simulations of their own lives, performing tasks with increasing precision.

---
Source: [Wired](https://www.wired.com/story/world-models-will-soon-be-inescapable-what-you-need-to-know/)  
Published by News Central Site: https://newscentral.site/niantic-shifts-focus-towards-developing-world-models-for-simulations/