The Information's TITV Nvidia GTC Preview, China’s SuperApp AI Advantage, SaaS’ AI Contradictions, Data Center Hacks
61 snips
Mar 16, 2026 Wayne Ma, NVIDIA reporter covering GTC and chip roadmaps. Ethan Choi, investor on the U.S.-China AI race and superapp strategies. Laura Bratton, AI reporter tracking SaaS regulatory risk disclosures. Ann Davis-Vaughn, AI infrastructure columnist on data center scaling tricks. They discuss Nvidia’s GTC previews and inference chips. They explore China’s superapp AI edge. They flag SaaS’ AI risk signals. They reveal industrial data center hacks.
AI Snips
Chapters
Transcript
Episode notes
Super Apps Make Agent AI More Powerful In China
- China's super apps bundle messaging, payments, e-commerce and physical services into one interface, enabling powerful agent experiences atop a single app.
- Ethan Choi says WeChat's integrated digital and physical infrastructure (warehousing, ride-hail, payments) makes an agent layer especially potent.
Open Weights Versus Closed Models Shapes The AI Race
- The US-China AI race splits along distribution and openness: China favors open-weight models and broad free distribution, the US favors closed models and heavyweight cloud players.
- Ethan highlights China's advantage in raw inputs like minerals, land and energy versus US advantage in data-center gigawatts.
Chinese Startups Innovate Around Hardware Limits
- Chinese model startups use hardware-constrained innovations like granular MoE and multi-token prediction to compete with US leaders.
- Ethan calls DeepSeek, Kimi and others formidable, often innovating from small teams and research labs to squeeze performance out of limited hardware.