At Autodesk University 2026 in Las Vegas, Haresh Khoobchandani, Vice President for Asia Pacific and Japan at Autodesk, argued that the next frontier for AI is physical, not digital, in a conversation building on a three-part LinkedIn series adapted from a public lecture he delivered in Singapore. His advice on a CEO panel was to "Stop talking about AI. Start talking about the problems you want to solve." A former flight steward with Singapore Airlines, he spent 19 years at Microsoft and the past eight at Autodesk, and frames AI as one layer in a broader system of data, workflows, and human decisions. Autodesk's strategy spans Design, Make, and Operate; historically fragmented data means applying AI to a disconnected environment yields garbage in, garbage out. The physical world raises the stakes, since a building cannot simply be generated and discarded, a bridge has to withstand real forces, and a factory has to work, making context the harder problem. The next phase may be defined less by bigger models and more by models that understand physical environments, reinforced by investments and partnerships like Autodesk's with World Labs, the venture led by Fei-Fei Li. Done right, data becomes a living asset that outlasts its creators, so institutional knowledge survives employee turnover.