Baidu Establishes Digital Human Innovation Business Unit
Baidu has undergone another restructuring, this time merging its commercial and e-commerce divisions while spinning off the digital human business into an independent unit. If memory serves, the concept of digital humans has been elevated to mythical status since the metaverse era, cooled off several times in between, and is now being revived and established as a separate entity, riding the wave of AI large models. The organizational reshuffling of major tech companies sometimes occurs more freq
Analysis
Baidu has restructured again, merging its commercial and e-commerce divisions while spinning off the digital human business into an independent department. If memory serves, digital humans have been hyped since the metaverse era, experienced several cooling-off periods, and are now being revived as a standalone entity, buoyed by the AI large model wave. The organizational reshuffling of major tech companies sometimes happens more frequently than fashion weeks—merging today, splitting tomorrow, all under the banner of "dynamic optimization based on market trends." But establishing a dedicated department for digital humans in 2026 exudes a strong mix of "trend-following" and "anxiety." Concepts riding the trend seem to secure resources and legitimacy more easily than concrete business ventures. Once a department is created, the story told in presentations seems more compelling, but in practice, beyond more lifelike virtual sales hosts and more mechanical customer service, what irreplaceable problem have we actually solved? Even Baidu’s own employees might pause to consider.
Almost simultaneously, across the ocean, Jensen Huang—currently the hottest tech leader on Earth—skipped glamorous conference rooms and instead met with the heads of two South Korean gaming companies in a Seoul Gangnam internet café. The setting itself was filled with dramatic tension. He discussed "physical AI" and humanoid robots with the head of KRAFTON and talked about game AI with representatives from NCSOFT, even live-streaming the encounter. The commercial banter was palpable through the screen. Huang is a true marketing master who understands the power of "scene." Discussing games and AI in an internet café creates more buzz, feels more grounded, and sends a clearer signal to global gamers and investors than any meeting room could: Look, we’re not talking about abstract algorithms—we’re engaging with the world you immerse yourselves in daily.
But amid this flashy "pop-in" and live-stream, we need to pour some cold water. Game AI is an old topic, from NPC behavior trees to procedural generation, and now to driving character dialogue with large models—progress is real. However, did Jensen Huang dive in personally just for the sake of in-game intelligence? The real intent likely lies elsewhere. KRAFTON has long collaborated on AI feature development, and NCSOFT is a seasoned MMORPG giant. The massive player behavior data and complex virtual world operational rules they’ve accumulated are the true treasures. Nvidia’s GPUs are the cornerstone of computing power, but raw power alone isn’t enough. They need these top-tier gaming companies’ scenarios to train and validate their AI models—not just for in-game AI, but for the general intelligence driving "physical AI" and robotics. Game worlds, as the most complex and dynamic simulated environments, have become perfect testing grounds for AI evolution. Huang’s café meeting is essentially a future-oriented resource-claiming exercise. Gaming has never been just entertainment; it’s a shortcut and data fuel for a broader AI vision.
Viewing Baidu’s organizational restructuring alongside Jensen Huang’s café diplomacy is particularly intriguing. One is about "interior renovation," elevating the symbolic "digital human" business to a standalone position; the other is about "extreme scene marketing" to consolidate the AI computing empire’s ecosystem. Both expose the deeper logic of the current AI competition: With no consensus yet on breakthroughs in underlying technologies (such as AGI), all players are scrambling to seize "application scenarios" and "data flywheels." Baidu is choosing to build digital humans as an independent stronghold within its existing mobile ecosystem, attempting to reignite the narrative on the ruins of the metaverse with large models. Nvidia, on the other hand, is more aggressive—not content to remain merely an arms dealer, it’s jumping into the trenches itself, binding with the top gaming battalions to ensure its computing ammunition lands on the most critical, data-rich battlefields.
So, let’s stop cheering for the mere establishment of new departments or celebrity photo ops. These moves are largely strategic positioning in response to competitive pressure, narrative supplements for the capital market, and inevitable noise in the pre-implementation phase of technology. Digital humans have become independent, but their soul (AI) remains universal; gaming collaborations have deepened, but the goals have long surpassed gaming itself. When the noise fades, we’ll see clearly which actions truly created new value and which merely slapped more dazzling new labels on old wine. In this sense, whether it’s Baidu’s "Digital Human Department" or Jensen Huang’s "café talks," they are merely two symbolically charged episodes in the long AI march. The real show has yet to begin.
Disclaimer: The above content is generated by AI and is for reference only.