Home Latest Feeds Technology News Nvidia’s newest AI tech can generate 3D models from prompts

Nvidia’s newest AI tech can generate 3D models from prompts

Nvidia’s newest AI tech can generate 3D models from prompts


Key Takeaways

  • Nvidia’s Latte3D AI model can create 3D models from text prompts, revolutionizing how developers work.
  • It can be trained to generate specific styles of 3D art, making it versatile and valuable for various industries.
  • The technology has the potential to lower the barrier of entry for independent game developers, saving time and resources.

AI has come a long way in a short amount of time. It feels like every month, AI can create better images, voice clones, and descriptions of concepts than it did the month before. However, despite all of its progress, AI hasn’t made much of a jump in generating 3D art. Nvidia is set to change that by announcing “Latte3D,” an AI model that can generate 3D models from a text prompt.


Nvidia GeForce RTX 5000 series: All of the rumors so far

Interested in the new Blackwell GPUs? Here’s everything you need to know about them

Nvidia’s new Latte3D AI model

As announced on the Nvidia blog, the graphics cards giant has been hard at work making an AI model that can take in text inputs and turn them into a 3D model. So far, AI has been mostly limited to 2D images and videos, but this development could boost how AI is used in the video game industry. If a developer wants to add a 3D model of something to their game, they can ask Latte3D to generate something and it will create a 3D asset for them to use.

What’s interesting about this technology is that it can be trained further by a developer to generate 3D models of a specific style or type. Nvidia listed a few examples where Latte3D was trained on two different inputs; a nature-based model would produce art based on natural elements while training it on household items would produce results that match what you’d see around a typical home.

Sanja Fidler, the vice president of AI research at NVIDIA, was very proud of the progress made:

“A year ago, it took an hour for AI models to generate 3D visuals of this quality — and the current state of the art is now around 10 to 12 seconds “We can now produce results an order of magnitude faster, putting near-real-time text-to-3D generation within reach for creators across industries.”

A potential lifeline for independent developers

A pair of spectacles balanced on a laptop with the screen showing code

Latte3D has a huge amount of potential for revolutionizing how independent developers work. If the model becomes public use, developers can generate lots of 3D models very quickly without the need to make or find them themselves. As such, this technology has the potential to lower the barrier of entry for indies to create 3D games; while they’ll still need to learn how to properly import (and potentially animate) the models, it would save a lot of time that would otherwise be spent either modeling or finding the right fit for their game.



Please enter your comment!
Please enter your name here