Zhipu AI jumps above HK$1T Monday, GLM-5.2 launch sends shares up 42%

Zhipu AI just crossed an attention line that is hard to ignore: on Monday, its Hong Kong-listed market capitalisation surpassed HK$1 trillion (US$128 billion). Shares of the company, which trades under the name Knowledge Atlas Technology, then ripped higher, soaring as much as 42% on Monday morning to a peak of HK$2,980.

That is not just a shiny number. The move is tied directly to a concrete catalyst the market can point at: Zhipu’s release of its open-source GLM-5.2 model last week. In other words, investors are pricing in not only “a new model exists,” but that the model can help Zhipu compete head-to-head with its American rivals.

To understand why this matters, you have to look at how AI markets behave when a vendor can credibly claim both capability and distribution. An open-source release lowers the barrier for developers and teams to try the model, build on it, and validate performance in real workflows. That can accelerate adoption because the tech can be inspected, tested, and integrated without waiting for a proprietary access arrangement. When the product is also a “rivalry” weapon, the effect can be immediate: the company signals speed, transparency, and momentum, while the market updates valuation expectations.

Zhipu’s particular bet is also about timing. The SCMP piece frames the jump as “investor optimism” tied to Zhipu going “head-to-head” with American rivals. That sets up a classic capital market dynamic: when investors believe an AI company is no longer playing catch-up, they tend to compress the perceived gap between “promising research” and “commercially relevant competition.” The 42% spike to HK$2,980 on Monday morning suggests the market treated GLM-5.2 as more than incremental news. It treated it as a repositioning.

There is also a listing-level angle. Because Zhipu trades in Hong Kong under the name Knowledge Atlas Technology, the rally is happening on a public market with its own shareholder base and liquidity patterns. When a Hong Kong-listed AI company can credibly claim global competitive relevance, the upside is not confined to private funding circles. Instead, valuation can re-rate quickly through public trading, and that can amplify visibility for both customers and talent.

The other non-obvious impact is how open-source model releases can tighten competitive feedback loops. Once a model like GLM-5.2 is released, competitors are pressured on multiple fronts at once: they face scrutiny from developers who compare versions, they face urgency to respond with their own improvements, and they face reputational pressure if they look slower. That is partly why investors react sharply. The market is not only buying the current model. It is buying the implied pace of iteration and the likelihood that Zhipu can keep up in an arena dominated by US-based leaders.

For executives and boards at AI companies, the strategic stakes are straightforward: market cap is not a measurement of “who has the best research.” It is a measurement of who the market believes is winning the next cycle of adoption, partnerships, and mindshare. When Zhipu can translate GLM-5.2 into a HK$1 trillion moment and a 42% share surge, it sends a signal that capital is willing to reward companies that ship open, developer-accessible models with clear competitive intent.

This is where the second-order implications start to matter. If Zhipu’s GLM-5.2 release becomes a proving ground for developer use, it can influence procurement and integration decisions, which in turn can feed back into revenues and further reinforce the story investors are already telling. Even if the initial headline effect is price-driven, open-source distribution can turn that narrative into traction. And once a company gets that traction narrative, peers often feel forced to accelerate their own release schedules or rethink how they compete, because investors will not wait forever for “eventually.”

In short, Monday’s crossing of HK$1 trillion is the headline, but the underlying signal is the escalation: Zhipu is using GLM-5.2 as a competitive instrument against American rivals, and the public market is reacting in real time. If you run a similar AI business, the takeaway is simple, and it is not gentle: the next model release can move your stock, but only if the market believes it changes your competitive trajectory.