Jack Ma’s public reappearance in a Hangzhou classroom this week was more than a hometown photo op. On March 3, Ma met with principals and teachers at Yungu School alongside Alibaba and Ant Group’s senior management — including Joseph Tsai, CEO Wu Yongming, e‑commerce chief Jiang Fan and Ant’s leadership — to discuss how artificial intelligence will reshape education and society. Ma warned that AI’s arrival is rapid and disruptive, and he framed teenagers as the generation with the greatest potential to adapt and benefit.
The visit underlines two things: China’s big tech firms are increasingly aligning internal strategy around AI, and they see education — from curriculum to talent pipelines — as a strategic front. By bringing core executives to a school, Alibaba signals a desire to shape public narratives about AI and to cultivate future users and builders at an early age. It is also a reminder that Chinese tech conglomerates are marrying corporate planning with social messaging in ways that domestic audiences will read as both philanthropic and strategic.
Meanwhile, the economics of generative media are shifting fast. ByteDance’s cloud arm published pricing for Seedance 2.0, a model for generating or editing video. The list shows two tiers: 28 RMB per million tokens when video input is included (editing), and 46 RMB per million tokens for pure generation. ByteDance says producing a 15‑second clip consumes roughly 308,880 tokens; at the higher rate that amounts to about 15 RMB per clip — roughly 1 RMB per second (about $0.14/s), a price point that makes short generative video trivially cheap for many commercial uses.
That drop in cost accelerates creative experimentation but also sharpens policy challenges. Cheap, high‑quality video generation lowers barriers for marketers, social creators and film editors, yet it also makes deepfakes and disinformation easier to produce at scale. The two price brackets — cheaper for editing, costlier for pure generation — reflect different compute intensities, but both are moving towards commoditization of synthetic video.
Capital and open‑source activity around AI remain vigorous. Deep Intelligence (DIP) announced a $40 million funding round from returning institutional investors, demonstrating continued private capital appetite for Chinese AI startups. Separately, Lingqu OS, a full‑stack, cross‑platform robotics software framework focused on bipedal locomotion and reinforcement‑learning training pipelines, was released as an open‑source Alpha, signaling increased investment into embodied AI and domestic robotics toolchains.
Investor faith in AI hardware also shows no signs of cooling. Liao Kaiyuan, once among Tesla’s largest individual shareholders, disclosed a purchase of 1 million Nvidia shares and said he will continue buying, calling AI “only beginning” and reiterating beliefs that energy and embodied AI businesses remain underpriced. On the global reputational front, reactions to AI firms’ ties with governments reappeared in the headlines: an OpenAI‑US Department of Defense collaboration prompted a spike in ChatGPT mobile app uninstalls in the United States, an indicator of consumer unease when commercial AI links to military projects.
Taken together, these events sketch an ecosystem moving from experimentation toward scale. Tech giants are coordinating education outreach and product launches, startups are closing meaningful rounds, open‑source projects are seeding the next generation of robotics, and investors are doubling down on AI infrastructure. The result is faster diffusion of capability — and a faster arrival of the regulatory and ethical dilemmas that follow.