Shares of Knowledge Atlas Technology JSC, widely known as Zhipu, edged higher on their Hong Kong trading debut on Thursday, as the Beijing-based artificial intelligence firm became the first of China’s so-called “AI tigers” to test public markets.

The stock rose as much as 15% above its offer price of 116.20 Hong Kong dollars ($15), after the company raised about $558 million in the offering. Roughly 37.4 million shares were sold, valuing the startup at around HK$4.3 billion and placing the listing among the more sizable AI flotations globally in recent years.

While the first-day gains were measured rather than explosive, the debut was closely watched by investors, policymakers, and competitors. Zhipu’s listing is being read as a referendum on whether Chinese large language model developers — operating under tightening U.S. export controls and rising geopolitical risk — can still attract capital, scale globally, and commercialize frontier AI technologies.

Founded in 2019 by researchers from one of China’s top universities, Zhipu has grown into a flagship player in Beijing’s artificial intelligence strategy. It is regarded as one of the country’s “AI tigers,” a label used to describe a small group of startups developing large language models intended to rival Western leaders such as OpenAI and Anthropic.

Unlike many consumer-facing AI firms, Zhipu’s rise has been closely intertwined with state priorities. The company is widely viewed as strongly backed by Beijing, both financially and politically, and sits at the center of China’s effort to reduce reliance on foreign AI technology while accelerating domestic innovation.

Its IPO, therefore, carries significance beyond the company itself. It represents the first time a major Chinese large language model developer has gone public via an initial offering, setting a precedent for peers that have so far relied on private funding rounds, state-linked investors, and strategic partnerships.

Among those peers is DeepSeek, another member of the “AI tiger” group, which rattled markets early last year with the release of one of its models. That launch challenged assumptions that U.S. export controls would sharply slow China’s progress in advanced AI, triggering renewed debate in Washington and elsewhere over the effectiveness of technology restrictions.

Zhipu, though less visible internationally than DeepSeek, has earned recognition at the highest levels of the global AI industry. Last year, OpenAI described the company as a notable competitor operating on the “front line” of China’s race to lead in artificial intelligence — an acknowledgment that underscored both Zhipu’s technical progress and the intensifying rivalry between U.S. and Chinese AI ecosystems.

The company has also pursued an unusually broad overseas footprint for a Chinese AI firm operating under sanctions. Zhipu reportedly maintains offices in the United Kingdom, Singapore, and Malaysia, alongside operations across the Middle East. It has established joint “innovation center” projects in Southeast Asia, including Indonesia and Vietnam, aligning with China’s broader push to deepen digital and technological ties across emerging markets.

Those international ambitions have advanced despite mounting regulatory obstacles. In January last year, the U.S. Commerce Department placed Zhipu on its Entity List, citing allegations that the firm was working with China’s military. The designation sharply restricted Zhipu’s access to advanced U.S. semiconductors, AI tooling, and certain forms of technical collaboration.

Like other Chinese AI developers, Zhipu has had to contend with U.S. controls limiting access to high-end chips and expertise needed to train and run large-scale models. These constraints have raised costs, complicated supply chains, and forced firms to rely more heavily on domestic alternatives, which often lag the most advanced foreign hardware.

Still, Zhipu has continued to invest heavily in model development. According to its prospectus, about 70% of the IPO proceeds will be directed toward research and development of its general-purpose large AI models, underscoring a strategy that prioritizes technical capability over near-term profitability.

The company reported revenue of 312.4 million yuan in 2024, a figure that highlights both rapid growth and the early stage of monetization for large language models in China. Industry-wide, revenue generation remains uneven, with firms experimenting across enterprise services, cloud deployments, customized models, and government contracts.

Zhipu’s public debut also comes at a moment when Chinese capital markets are selectively reopening to technology firms aligned with national priorities. In recent months, several AI chipmakers and semiconductor-related companies have listed, suggesting regulatory support for areas viewed as strategically vital to economic and national security.

Market participants see Zhipu’s IPO as a test case for whether investor confidence can extend beyond hardware into software-driven AI models, particularly those facing external sanctions and internal policy scrutiny.

The listing may soon be followed by others. Rival Chinese AI startup MiniMax is expected to launch its own offering on Friday, after submitting a confidential filing last year. A successful debut could signal a broader wave of public listings from China’s AI sector, while a weak showing could reinforce doubts about the sector’s commercial readiness and risk profile.

However, Zhipu’s restrained but positive first-day performance points to cautious optimism. Investors appear to be weighing the company’s strategic importance, state backing, and technical progress against persistent uncertainties surrounding geopolitics, export controls, global competition, and the high capital demands of training large-scale AI models.

In that sense, Zhipu’s IPO is not just about one company’s market debut. It is a live experiment in how far China’s AI ambitions can travel in public markets at a time when technology, capital, and geopolitics are increasingly inseparable.