The AI company known as "the first global large model stock," KNOWLEDGE ATLAS (Stock Code: 02513.HK), experienced a dramatic surge in its share price on Friday, February 20, skyrocketing over 42% in a single day. This propelled the company's market capitalization past the HKD 320 billion mark.
However, the celebration was short-lived. The very next day, the high-profile AI firm issued a letter of apology directly to its developer community. It openly admitted that its newly launched flagship large model, GLM-5, was facing significant operational challenges and difficulties following its release.
The domestic AI company KNOWLEDGE ATLAS presented a tale of two extremes on the same day. On one hand, its stock price surged wildly. This was driven by international recognition for its new-generation GLM-5 model technology, the fact that its core product sold out despite a 30% price increase, and a scarcity of shares available for trading. This led to the 42% single-day jump in its Hong Kong stock, with the market betting that its "technological leadership" could translate into commercial victory.
On the other hand, its operations stumbled. After the new model's launch, issues such as opaque billing rules, delayed server capacity expansion, and a problematic upgrade experience left many paying developers unable to use the service or facing incorrect charges. The company quickly issued an apology letter the following day, acknowledging its shortcomings and offering compensation including full refunds and subscription extensions.
The company's essence is that of a technically-focused "AI model supermarket" founded by top graduates from Tsinghua University. While its revenue is growing rapidly, it is also investing heavily in research and development, reporting an annual loss of CNY 3 billion, essentially racing ahead by burning cash. The risks lie in the fast pace of technological obsolescence, heavy reliance on imported chips, and dependence on top-tier talent.
In summary, its technology is strong enough to drive the market wild, but its operations are poor enough to frustrate users. The stock price frenzy is a bet on the company's ability to quickly address its commercial shortcomings.
The first question is: why did the stock price rise 42% in one day? It wasn't due to a perfect product, but rather a perception of its potential for dominance. The technology is genuinely advanced; the new GLM-5 model received recognition from overseas developer communities, suggesting it is approaching top international standards. This signaled to the market that China can produce world-class AI, even under chip restrictions, using domestic alternatives, which greatly expanded the imagination for its potential.
It dared to raise prices and still sold out; a new programming package priced 30% higher sold out immediately upon release. This indicates that its technology is good enough for programmers to deem it "worth the price," demonstrating real capability rather than competing on low cost. Additionally, with a recent IPO, most shares are still in a lock-up period, leaving very few shares available for free trading. High demand from investors seeing potential, coupled with limited supply, pushed the price sharply higher. Thus, the stock surge was a market vote of confidence in its "technical strength" and "profit potential."
The second question is: why issue an apology letter immediately after the price surge? Because the reality fell short of the ideal. While the technology is advanced, the company stumbled in its execution. The apology letter cited three main issues. The rules were unclear; the new model consumes significant computing power, leading the company to implement an undisclosed "peak usage multiplier" rule, causing users' account balances to drop unexpectedly and creating frustration. The servers were unprepared; the model's popularity, combined with scalpers grabbing resources, led to insufficient server capacity, leaving paying users unable to access the service or experiencing slowdowns during peak times. The upgrade experience was problematic; the interface for existing users to upgrade was confusing, leading some to make unintended selections and incur unwanted charges. Simply put, the product is good, but the service delivery was lacking, alienating its earliest supporters.
What is the background of this company? According to its prospectus, it is essentially an "AI model supermarket" founded by a group of top Tsinghua University graduates. It doesn't sell physical goods but various "AI brains" capable of writing articles, creating images, or assisting with phone operations. It generates revenue by leasing these "AI brains" to enterprises, charging based on usage, or selling them outright to major clients like banks and governments for on-premises installation to ensure data security.
Financially, it follows a typical "high-growth, cash-burning" model. Revenue is growing rapidly, reaching hundreds of millions annually, but losses are even larger, amounting to billions per year. Funds are heavily spent on expensive graphics cards and retaining scientific talent. The company maintains a good gross margin, indicating the underlying business model is profitable, but it currently reinvests all earnings and raised capital into research and development to capture market share.
The primary risks include the rapid pace of technological change in AI, where falling behind can mean quick obsolescence. The company remains a significant cash burner, requiring continuous funding. It is highly dependent on its core team of top scientists; losing them would be a major setback. It also faces the risk of being constrained by reliance on chips from companies like NVIDIA, and it operates in a highly competitive landscape against tech giants like Baidu, Alibaba, and Tencent.
In conclusion, the situation can be likened to a gifted student who aces an exam but makes a careless mistake on homework, leading to a reprimand. The market is divided. Optimists believe the student's core talent remains, and minor errors can be corrected, viewing any stock price pullback as a potential buying opportunity. Pessimists see the operational mishap as revealing a lack of practical savvy, questioning how it can compete against established giants and suggesting the current high valuation may already be overly optimistic, representing a potential bubble.
The plain-language summary is that KNOWLEDGE ATLAS currently possesses technology impressive enough to earn applause but operations frustrating enough to provoke criticism. The stock's frenzy is a gamble on its ability to successfully convert its "technological advantage" into a "commercial victory." The apology letter serves as a dose of reality for this high-stakes bet.