Zhipu AI’s Market Debut and Stock Performance
Shares of Knowledge Atlas Technology JSC, widely known as Zhipu AI, experienced a dramatic surge on Wednesday, climbing as much as 35% following the release of its inaugural earnings report as a publicly traded company. The Beijing-based firm ultimately closed the session 31.94% higher, marking a significant market reaction to its first full-year financial results since its Hong Kong listing in January. The initial public offering, which raised $558 million, had positioned Zhipu as the first major pure-play AI model company from China—and arguably globally—to list on a major exchange, drawing intense scrutiny from investors keen to gauge the commercial viability of China’s homegrown AI champions.
Earnings Report: Growth Amid Persistent Losses
Zhipu’s report, released on Tuesday, revealed robust top-line growth. The company’s revenue for 2025 rose approximately 132% year-over-year to 724 million yuan. However, this figure fell slightly short of the mean analyst estimate of 760 million yuan, according to a Reuters poll. The revenue miss was juxtaposed against a continued substantial net adjusted loss, which widened by 29.1% from the previous year to 3.18 billion yuan. The company attributed this increased deficit to heightened research and development (R&D) expenditures, a common theme among AI firms investing heavily in model training and talent acquisition. This financial picture underscores the classic growth-over-profitability trajectory common in frontier technology sectors, where massive upfront R&D costs precede potential scale.
Technological Position and Market Reach
Founded in 2019 by researchers from Tsinghua University, Zhipu has rapidly ascended to be considered one of China’s premier “AI tigers”—startups building large language models (LLMs) to compete with global leaders like OpenAI and Anthropic. The company highlighted the recent release of its GLM-5 model, stating it matched key performance metrics of leading U.S. rivals. Beyond core models, Zhipu has been expanding its portfolio of AI agent offerings. The firm reported that over 4 million small and medium-sized enterprises and developers have collaborated with its products, which are now accessible in 218 countries and regions. A significant driver of recent usage was the nationwide adoption frenzy for its open-source AI agent, OpenClaw, which pushed token usage to unprecedented levels. Tokens, the fundamental units of text processed by LLMs, serve as a direct metric for computing demand and associated operational costs.
Geopolitical Headwinds and Strategic Alignment
Zhipu’s trajectory is inextricably linked to China’s national AI strategy and the broader U.S.-China tech rivalry. During the earnings call, CEO Zhang Peng emphasized the company’s acceleration in using domestically produced Chinese chips to address a “significant rise in computing demand” since February. This pivot directly aligns with Beijing’s strategic push for semiconductor self-sufficiency. The CEO’s remarks are particularly notable given the firm’s prior constraints; U.S. export restrictions have historically limited Zhipu’s access to advanced semiconductor technology and expertise crucial for training cutting-edge AI models. Compounding these challenges, Zhipu was placed on the U.S. Commerce Department’s Entity List in January 2024 over alleged links to China’s military, a designation that complicates its access to U.S. technology and software.
Broader Implications for China’s AI Sector
Zhipu’s public market performance and operational data serve as a critical barometer for the health and momentum of China’s entire AI ecosystem. The strong stock reaction, even with a minor revenue miss, suggests investor optimism about its long-term prospects and the vast domestic market opportunity. The performance of another Chinese AI rival, MiniMax, which also listed in Hong Kong in January and saw its shares rise about 16% on the same day, indicates a broader positive sentiment towards the cohort. As these companies navigate the dual challenges of scaling commercially and operating under technological restrictions, their ability to innovate with domestic supply chains and capture enterprise and developer demand will be pivotal in determining China’s competitive standing in the global AI race.
Sopa Images | Lightrocket | Getty Images
This article is based on information from CNBC and incorporates data cited from Zhipu AI’s earnings report and analyst polls from Reuters. All original links and HTML elements have been preserved.
