OpenClaw and the "Little Lobster" Fever: A Qualitative Study of China's 2026 AI-Agent Adoption Wave

Abstract

In March 2026, thousands of people queued outside technology campuses in Shenzhen, Beijing, and Hangzhou to have an open-source agent framework called OpenClaw installed on their machines. Chinese users renamed the tool the "Little Lobster," and the act of running it became "lobster farming." This paper treats that episode as a case worth explaining rather than a curiosity to report. Using qualitative document analysis of policy texts, corporate announcements, community discussion, and reported adopter accounts, it asks who joined the wave, why the Chinese deployment of agent software diverged so sharply from its Western counterpart, and what the episode reveals about how a large economy governs a fast-moving technology. The analysis finds a single social driver beneath four visible tensions, a policy structure that pairs central-level red lines with local-level competition to attract investment, and a corporate value cycle that turns individual anxiety into platform revenue. I introduce three interpretive concepts, radical pragmatism, algorithmic survivalism, and digital narcotization, to name the mechanisms at work, and I argue that the phenomenon is best read as a stress test that China is running on behalf of the wider world, at a cost borne mainly by ordinary users. The paper closes by sketching the likely path from an improvised, command-line era toward regulated, platform-native agent services.