Luma's Leap into AI-Driven 3D Modeling

Luma’s Leap into AI-Driven 3D Modeling

Jan. 12, 2024. 2 mins. read.

Interactions

Revolutionizing 3D Capture with Smartphone Technology
A few years back, Alex Yu, an AI researcher at UC Berkeley, and Amit Jain, an Apple multimedia specialist, embarked on an ambitious journey with Luma. Their goal was simple yet groundbreaking: enable people to capture 3D objects using just their smartphones. Fast forward to today, and their app has amassed over two million users, signaling a successful first chapter in their innovative venture.

Evolving with Generative AI
As generative AI technology makes waves, Luma is not resting on its laurels. The company recently announced a bold new direction, utilizing a massive compute cluster of around 3,000 Nvidia A100 GPUs. This upgrade aims to train cutting-edge AI models that can interpret, illustrate, and eventually interact with the world in unprecedented ways. Luma’s initial focus is crafting 3D models from textual descriptions, as demonstrated by their model ‘Genie’ on Discord.

Addressing the Uncanny Valley
Yu and Jain recognize the challenges in current-gen generative AI, notably the ‘uncanny valley’ issue, where AI-generated models often lack realism. To combat this, Luma is developing next-generation AI models to enhance multimodal intelligence, combining vision with other senses to create more life-like, realistic 3D models.

Securing Funding for Future Innovation
To fuel these ambitious plans, Luma has successfully raised $43 million in a Series B funding round. With this boost, Luma’s valuation soars between $200 and $300 million, with a total war chest now exceeding $70 million. This financial injection will accelerate Luma’s roadmap towards developing advanced photorealistic technologies in an intuitive app.

Navigating a Competitive Landscape
Luma is not alone in this race. The field is bustling with competitors like 3DFY, Scenario, and Hypothetic, each vying for dominance in 3D model generation. But Luma’s unique selling point lies in its focus on fidelity. Yu believes that their approach to generating coherent and usable models in fewer attempts sets them apart.

The Road Ahead
While an improved version of Genie is already out, more sophisticated AI models are on the horizon. Luma is gearing up for this challenge, planning to double its workforce and assembling a formidable server cluster. In the dynamic world of AI and 3D modeling, Luma is poised to make significant strides, potentially revolutionizing how we perceive and interact with digital 3D objects.

Interesting story? Please click on the ? button below!

SOURCE: Luma raises $43M to build AI that crafts 3D models | TechCrunch