The brain drain from tech giants toward the agile structures of startups continues unabated as 2026 emerges as a year of structural realignment for the global Artificial Intelligence industry. At the heart of this shift is Andrew Dai, one of the most prominent researchers to depart Google DeepMind, who chose the stage of the JPMorgan Global China Summit to discuss the progress of his new venture, Elorian.
The Transition from Recognition to Reasoning
Elorian is not merely another computer vision firm. According to Dai, the current generation of AI models, while impressive at generating or describing images, lacks what scientists call "visual reasoning." While a traditional model can identify a hammer and a nail in a photograph, Elorian seeks to build systems that understand the physical interaction between them, their spatial utility, and the consequences of an action.
This approach represents the next great frontier in AI. The ability of a system to "think" visually is essential for developing advanced robotics, autonomy in complex environments, and design tools that grasp real-world constraints. Dai emphasizes that Elorian focuses on building models that do not rely solely on statistical probabilities of pixel occurrences but on a deeper understanding of causality.
Funding and Growth Strategy
Despite the investment landscape in 2026 being more discerning compared to the "explosion" of 2023-2024, Elorian appears to be attracting significant interest. Dai’s presence at the JPMorgan China summit was no coincidence. Asia, and particularly the manufacturing and robotics ecosystem in the region, serves as a prime application field for Elorian’s technology. Dai revealed that the company is in advanced discussions for new funding rounds, targeting investors who understand the long-term horizon of "deep tech."
Elorian’s hiring strategy is equally aggressive. The firm aims to attract top-tier talent from academia and Big Tech R&D departments, offering something Google or Meta now struggle to guarantee: freedom from corporate bureaucracy and the ability to have a direct impact on the final product. "At Elorian, every line of code counts," Dai noted, highlighting the importance of agility in a rapidly evolving market.
The Geopolitical Factor and the China Market
The choice of the JPMorgan Global China Summit as a platform for these updates underscores the complex relationship between Western innovation and the Chinese market. While chip export restrictions and geopolitical tensions remain at the forefront, the need for research collaboration and access to massive datasets from Asian industries makes the region an inevitable destination for any serious AI startup. Dai seems to be navigating a fine line, seeking global reach while maintaining his company’s technological foundation within Western governance standards.
"The future of AI is not just about talking to machines, but about enabling them to understand and interact with the physical world in a logical and predictable way," Dai stated.
In conclusion, Elorian represents a new generation of companies that are not satisfied with just "generative" AI but are aiming for "intelligent" AI. Its success will depend on whether Dai’s team can translate the theoretical breakthroughs of DeepMind into practical, scalable solutions for the global market.
