Elorian, a new startup co-founded by former Google DeepMind researcher Andrew Dai, has emerged from stealth with $55 million in funding to tackle one of AI’s biggest hurdles: physical reasoning. Backed by a $300 million valuation from investors like Nvidia and Menlo Ventures, the Palo Alto firm is moving beyond simple image generation to build models that actually understand the mechanics of the real world. Dai noted that current systems fall short in visual reasoning, stating, “It’s much easier to generate something that looks great, but it’s a lot harder to reason about it.”
As reported by Bloomberg, Elorian plans to release its first reasoning model within the next year and is already in discussions with potential customers across industries such as architecture, automotive, and robotics. The company is developing its technology using open-source models while considering a hybrid approach that includes proprietary systems. Dai emphasised the importance of improving how AI interprets physical-world data, adding, “You actually need to design the physical thing—and that design lives in the physical world.”
