New Stock News | Diginex (00600) ends its IPO subscription and the oversubscription amount reaches 4.84 billion HKD, oversubscribed 15.4 times.

Artificial intelligence reasoning system chip supplier Aixinyuanzhi (00600) launched its IPO from January 30 to February 5. Aixinyuanzhi plans to issue 100 million H-shares, with 10% to be publicly offered in Hong Kong at a price of HK$28.2 per share, raising HK$2.96 billion. As of the afternoon of February 4, Aixinyuanzhi recorded HK$48.4 billion in oversubscription, with the public offering raising HK$2.96 billion, oversubscribed by 15.4 times. According to the IPO plan, Aixinyuanzhi will have a board lot size of 100 shares, with an entry fee of HK$2848.4. It is expected to be listed for trading on February 10. CICC, GUOTAI JUNAN I, and BOCOM INTL are the joint sponsors. In recent years, Aixinyuanzhi has introduced investments from Chinese tech companies Meituan (03690) and Tencent (00700), as well as mainland venture capital firms Qiming Venture Partners and Yuanhe Puhua. Aixinyuanzhi has also brought in cornerstone investors such as Weishi Semiconductor Hong Kong Limited, New Horse International Limited, and Minshan Capital Limited, with total investments of nearly $190 million. According to the prospectus, the core technology of Aixinyuanzhi's SoC products is the Aixintongyuan (Axera Neutron) mixed-precision neural network processor (NPU), which is a specialized processing architecture that achieves good AI inference performance through advanced mixed-precision computation. This technology is crucial for deploying quantized models and enabling AI inference on edge and terminal devices. In particular, the expanding intelligent automotive industry is driving the accelerated adoption of this technology. Complementing the mixed-precision NPU is Aixinyuanzhi's Aixinzhimou (Axera Proton) AI-ISP, the world's first commercially available AI image signal processor at scale. Aixinzhimou AI-ISP is an advanced image signal processing engine that optimizes visual data at a pixel level in real-time, ensuring high-quality imaging even under strict conditions. Combined with the company's other technologies, such as the Pulsar2 toolchain and software development kit (SDK), these innovations address the basic "computing" needs of AI inference and create value through essential "perception" applications. According to data from Jacosol Consulting, Aixinyuanzhi is the largest provider of high-end visual edge AI inference chips globally by shipment volume in 2024. The same data source also indicates that Aixinyuanzhi is a leading player in the global visual edge AI inference chip market and is ranked among the top five in 2024. As of September 30, 2025, Aixinyuanzhi has independently developed five generations of SoCs and commercialized them, covering dozens of types, and has achieved mass production in visual terminal computing, intelligent automotive, and edge AI inference applications.