Knowledge Atlas surges 7% as founder pledges AGI focus over monetization

Tang Jie's "Touch High Plan" commits Knowledge Atlas to AGI development over near-term revenue, a bet the market rewarded with a 7% surge.

Knowledge Atlas (02513.HK) surged as much as 7.1% to HK$1,757 on Monday after co-founder Tang Jie published an internal letter pledging the company would prioritize artificial general intelligence over short-term application monetization.

"Not reaching the summit means failure," Tang Jie wrote in a nearly 4,000-word letter titled "The Giant Wave Has Arrived," outlining a two-year strategic initiative called the "Touch High Plan" focused on long-horizon task planning, autonomous agents and self-evolving AI systems.

The stock opened 4.9% higher and extended gains on turnover of HK$3.66 billion, with 2.1 million shares changing hands. The rally comes after the stock fell from above HK$1 trillion in market capitalization to about HK$730 billion following the July 8 lock-up expiration, mirroring the post-lockup slide at rival MiniMax.

Tang's letter is a deliberate attempt to frame Knowledge Atlas as an AGI company rather than an AI coding tool provider, a distinction that carries vastly different valuation multiples. OpenAI and Anthropic command trillion-dollar valuations based on future potential, while MiniMax was re-priced as a consumer AI app company after its lock-up expiration, with its market cap falling below HK$100 billion.

The Coding Trap

Knowledge Atlas's market capitalization once exceeded HK$1 trillion, driven largely by its coding capabilities. The company's MaaS platform reached 1.7 billion yuan in annual recurring revenue, a 60-fold increase over the past year, and its open-source GLM-5.2 model trails only Claude Opus 4.8 and GPT-5.5 on core benchmarks. But Tang barely mentioned coding in his letter, according to analysts who reviewed it.

"Once a story starts to be realized, it is no longer the future for capital markets," said a Hong Kong-based technology analyst who asked not to be named discussing private strategy. "Tang is trying to get ahead of the valuation question before the market forces one on him."

The concern is well-founded. MiniMax's stock plunged after its lock-up expiration in early July as the market applied traditional SaaS valuation metrics — ARR size, growth rate, user retention — to what had been priced as a frontier AI company. Its market cap fell below HK$100 billion. The broader backdrop compounded the pressure: global AI concept stocks faced a Q2 2026 selloff as expectations of Federal Reserve rate hikes rose and enterprise IT budgets contracted.

The AGI Bet

Tang's "Touch High Plan" commits the company to three milestones: long-horizon task planning spanning weeks and months, autonomous agent systems capable of self-directed collaboration, and self-evolving AI where models train other models. He framed the evolution as moving from "One Person Company" to "No People Company" — AI progressing from writing code for programmers to running entire organizations.

None of these capabilities have been commercialized at scale by any company globally. The strategy mirrors moves by OpenAI, which shifted its product focus to Operator and Deep Research after GPT-5, and Anthropic, whose updates this year center on Computer Use and agent loops. Tang cited Google DeepMind's "From AGI to ASI" report in his letter, arguing that superintelligence could emerge from continued compute scaling even without breakthroughs in model architecture.

Knowledge Atlas did not disclose the budget for the Touch High Plan or provide a timeline for commercialization. Tang stated the company will not pursue short-term application monetization, a stance that may test investor patience if technological breakthroughs take longer than expected. For investors, the bet is binary: either Knowledge Atlas delivers on its AGI roadmap and justifies a valuation comparable to OpenAI's reported $300 billion or Anthropic's $60 billion, or it follows MiniMax's path of valuation compression as the market shifts to revenue-based metrics. The next catalyst is the company's interim results, where investors will look for signs of R&D spending discipline and any early agent-related revenue.

This article is for informational purposes only and does not constitute investment advice.