According to Zhitong Finance APP, China Securities Co., Ltd. (ASX: 601066) has released a research report stating that overseas security discussions have not changed the high-frequency model iteration pace, and RSI is gradually becoming a consensus among leading model companies such as OpenAI and Anthropic. Domestically, Alibaba (09988) and Zhipu have completed approximately HKD 80 billion and USD 5 billion in financing respectively, with large-scale investment supporting model capability catch-up, while TENCENT (00700) has also increased overseas computing power procurement. On the commercialization front, overseas enterprise cooperation is deepening, with ChatGPT advertising annualized revenue run rate reaching USD 1 billion, and Muse driving broader consumer monetization pathways. Domestically, Agents are still expected to drive Token calls and a new round of ARR growth by increasing per-user task density. The overall domestic AI sector is trending upward, and attention is recommended for computing power services, domestic chips and super nodes, as well as B-end AI application vendors with scenario, data, and enterprise delivery capabilities. The main views of China Securities Co., Ltd. (ASX: 601066) are as follows:

Overseas: Model Capabilities Continue to Improve, Enterprise Deployment and Consumer Monetization Take Hold

Security discussions have not changed the high-frequency release pace of models. OpenAI and Anthropic have successively updated GPT-6 Sol/Luna, Opus 5.5, and GPT-6.1 Sol, with the latter approaching GPT-6 Astra in capability while costing only one-fifth as much. R&D automation is beginning to feed back into model iteration. As of August, Claude's human-supervised R&D work proportion rose to 26%. On the commercialization front, Anthropic has deepened cooperation with Salesforce and Bain to push Agents into enterprise processes. ChatGPT advertising annualized revenue run rate has reached USD 1 billion, and Muse has expanded personal task entry points, beginning to release consumer-side commercial value.

Domestic: Financing Supports Model Expansion, Agent Tasks Accelerate ARR Growth

Alibaba (09988) completed approximately HKD 80 billion in placement, Zhipu completed approximately USD 5 billion in equity and debt financing, and TENCENT (00700) also increased overseas computing power procurement. On the commercialization front, MiniMax's July Token consumption reached 20 times that of January, and Zhipu raised its year-end ARR target from USD 2.4 billion to USD 3 billion. Competition in office Agents and the Harness ecosystem are still evolving. The next stage of growth will depend more on effective task volume, usage retention, and paid conversion. Model capability improvements are expected to drive a new round of revenue expansion.

Market Review: Hardware and Overseas Software Recovery Lead

Using July 31 closing as the base period, as of September 30, the IGV and Philadelphia Semiconductor Index rose 12.6% and 11.6% respectively. A-share communications and electronics rose 4.7% and 3.2%, while computers and media fell 5.2% and 3.2%. On valuation, IGV's comprehensive PS-TTM was 8.59 times, while Shenwan computers and media were 2.80 times and 2.68 times respectively, down 10.1%, 21.6%, and 18.6% from the end of 2025. AI payment and data call growth in overseas software provide support for valuation recovery, while domestic application-side performance remains differentiated. Subsequent pricing will place greater emphasis on customer procurement, sustained usage, and profit contribution. Enterprises with clear AI revenue and delivery capabilities are more likely to receive valuation support.

Risk Warnings

(1) AI industry commercialization falling short of expectations: Currently, the commercialization models of AI products at various stages are still in the exploration phase. If the pace of advancement of products at each stage falls short of expectations, it may adversely affect the performance of related companies. (2) Market competition risk: Overseas AI vendors,凭借 first-mover advantages and strong technical accumulation, hold an advantageous position in competition. If domestic AI vendors' technology iteration falls short of expectations, their operating conditions may be affected. Meanwhile, many domestic enterprises have already invested in AI product R&D, and subsequent homogeneous competition risks may affect related companies' revenue. (3) Policy risk: The development of AI technology is directly affected by the policies and regulations of various countries. As AI penetrates various fields, governments may further introduce corresponding regulatory policies to regulate its development. If enterprises fail to adapt to and comply with relevant policies in a timely manner, they may face corresponding penalties or even be forced to adjust business strategies. In addition, policy uncertainty may also lead to errors in corporate strategic planning and investment decisions, increasing operational uncertainty. (4) Geopolitical risk: Amid fluctuations in the global geopolitical environment, especially US export restrictions on China may directly affect domestic enterprises' access to computing power chips, thereby affecting their product R&D and market competitiveness. At the same time, geopolitical risks may also cause AI products to face obstacles in expanding into overseas markets, affecting related companies' revenue conditions.