Descriptions:
Prof G Markets host Ed Elson is joined by Alice Horn (co-host of the China Decode Podcast) to break down a wave of Chinese AI model launches timed to Lunar New Year 2026. The releases include Alibaba’s Qwen 3.5 (a coding and agent-focused model reported to be five times faster than its predecessor) and Renbrain (a robotics perception model), ByteDance’s Seance 2.0 (video generation) and Duba 2.0 (deep reasoning), and ZepO’s GLM5 (agentic intelligence). Horn notes that across benchmarks, these models are matching or exceeding US competitors while costing 10 to 20 times less than equivalent tiers from OpenAI and Anthropic.
The conversation draws a meaningful use-case distinction: Chinese open-weight models are being downloaded locally and fine-tuned by engineers — Airbnb CEO Brian Chesky is cited as an example of an enterprise adopter using Chinese models for agentic chatbots due to cost and speed advantages — while US closed models retain an edge in high-precision reasoning tasks. Alice Horn argues against direct comparison, suggesting the two ecosystems are converging on different market segments rather than competing head-to-head.
The episode also includes a segment with Robin Brooks, senior fellow at Brookings and former Goldman Sachs chief FX strategist, examining Sweden’s formal inquiry into euro adoption as a signal of eroding dollar confidence — the euro has climbed 13% over the past year while the dollar has declined 9%. Together the segments paint a picture of accelerating geopolitical and technological decoupling with direct implications for AI infrastructure investment.
📺 Source: The Prof G Pod – Scott Galloway · Published February 19, 2026
🏷️ Format: News Analysis
