B[The Hang Seng Index and Hang Seng Technology Index both rose by about 1%, and pharmaceutical stocks and biotechnology stocks rose] On October 6, Hong Kong stocks closed higher, with the Hang Seng Index rising by 1% and the Hang Seng Technology Index rising by 0.94%. Zhipu rose more than 7%, and GLM-5.3 was listed on Amazon Bedrock. Pharmaceutical stocks and biotechnology stocks rose. CanSino Biologics rose by more than 14%, Biocytogen-B rose by more than 11%, Zai Lab rose by more than 9%, Insilico Intelligent, Sino Biopharmaceuticals, and Kangfang Biologics rose by more than 4%, 3SBio and Chi-Med Pharmaceuticals rose by more than 3%, and Hopewell Pharmaceuticals-B rose by about 3%. A[Zhipu: Raising US$5 billion in the form of zero interest + premium stock conversion to release long-term equity expectations] September 13th, on September 13, Zhipu (02513.HK) announced the completion of approximately US$5 billion in financing, including approximately US$2 billion in share placement and approximately US$3 billion in convertible bond issuance. This financing is mainly used for the next generation GLM model, fully self-training system and related computing infrastructure. According to the announcement, the placement price of Zhipu shares is HK$714 per share, a discount of approximately 9.96% to the closing price before the announcement. The placement shares account for approximately 4.50% of the enlarged issued share capital. The convertible bonds adopt a zero-coupon structure, with an issuance price of 100.5% of the principal and redemption of the principal upon maturity; the initial conversion price is HK$892.50 per share, a 25% premium over the placement price and a premium of approximately 12.55% over the closing price before the announcement. The combination of zero interest and premium issuance is an important signal for this round of financing. As the R&D budget and funding arrangements are further clarified, Zhipu's capital base for continuing to carry out training experiments, promote model iteration and improve the computing power system has been strengthened, adding chips to participate in the next stage of global cutting-edge model competition. B[Zhipu responds to the question of "knowledge only after training": The next generation base model is already advancing GLM-6.0, aiming at "self-evolution"] September 1st, at the 2026 semi-annual performance communication meeting held on the evening of August 31, an analyst raised the market's doubts about Zhipu's "knowledge only after training" and asked why the company has been relatively restrained in parameter scale recently. Tang Jie, founder of Zhipu, said that the next generation model will still expand the base size, and at the same time control the activation parameters to avoid the decline in inference speed and increase in costs. In terms of computing power, Zhipu revealed that it has achieved large-scale inference with 100,000-level domestic chips, and the cost of inference per Token has dropped by 80% from the beginning of the year. GLM‑5.3‑Flash is Zhipu’s first model that fully relies on domestic chip clusters to provide services under ultra-large-scale real traffic. The company said that compared with the initial baseline of the same hardware, the end-to-end service performance of this model has been improved by 3 times. Tang Jie summarized one of the future directions of GLM-6.0 as self-evolution, "This is a big challenge. The biggest problem is that the model can judge by itself when to stop and when to correct itself. This is the biggest problem, rather than simply training the model to a larger size. In future models, this is a key research focus." (every sutra)
— All loaded —
