The current overseas focus on AI infrastructure has shifted from cloud vendor spending to AIDC cloud computing centers.

According to the information from Zhitong Finance APP, Minsheng Securities released a research report stating that the current focus of AI infrastructure overseas has shifted from cloud vendor expenditures to AIDC cloud computing centers, with companies possessing GPU card first-mover advantages such as Oracle, CoreWeave, NEBIUS, Applied Digital, and Iris Energy gaining attention. Domestically, the bank believes that the AI narrative logic remains unchanged, with the focus being on the AI infrastructure side. It is recommended to pay attention to the AIDC chain: Runze Technology (300442.SZ), Runjian Co., Ltd. (002929.SZ); and optical module companies resonating both domestically and internationally: Dekoli (688205.SH), Shiji Photonics (688313.SH), Zhongji Xuchuang (300308.SZ), New Easyshine (300502.SZ), and Tianfu Communication (300394.SZ).

Key points from Minsheng Securities are as follows:

NVIDIA invests in UK AIDC, marking the start of comprehensive AI expansion overseas. According to Sina Finance news on September 18, 2025, NVIDIA will invest £500 million (equivalent to $6.83 billion) in Nscale and stated that this UK-based company will participate in artificial intelligence infrastructure construction projects worth up to £11 billion.

Oracle issues market guidance, expanding its cloud business. The tech giant Oracle issued performance guidance, expecting revenue for the second quarter of fiscal year 2026 to grow by 14%~16% year-over-year, with adjusted earnings per share projected between $1.61 and $1.65. This growth is mainly driven by strong performance in the cloud business, with first-quarter cloud infrastructure revenue reaching $33 billion, a year-over-year increase of 55%. The company’s remaining performance obligations (signed but not yet recognized revenue) amounted to $455 billion, surging 359% year-over-year, thanks to multi-billion-dollar contracts signed with companies like OpenAI. To support this growth, Oracle plans to increase capital expenditure to $35 billion in the current fiscal year, representing a 65% year-over-year increase.

NEBIUS secures nearly $20 billion order from Microsoft. NEBIUS reached a significant agreement with Microsoft Corporation, with the transaction amounting to nearly $20 billion. This deal will bring $17.4 billion to the company, while Microsoft retains the option to purchase an additional $2 billion in services until 2031, making the total potential value up to $19.4 billion. Under the agreement, NEBIUS Group will provide exclusive capacity for Microsoft at its new data center in Vineland, New Jersey, to be deployed in multiple phases throughout this year and next year, with services expected to commence later this year.

The bank believes that global investment in AI infrastructure is accelerating. According to the latest data estimates from Omdia, global spending on cloud infrastructure services reached $95.3 billion in the second quarter of 2025, a year-over-year increase of 22%. Market growth momentum remains robust, maintaining over 20% year-over-year growth for four consecutive quarters. In July this year, Microsoft

announced that its annual AI capital expenditure would reach $120 billion, significantly up from last year’s $88 billion, almost four times the $32 billion spent in 2023. Meta expects its total expenditure for 2025 to range between $66 billion and $72 billion, primarily for expanding server, network, and data center construction required for generative AI, with a lower limit of $64 billion. Meta’s Chief Financial Officer Susan Li stated that the company anticipates further increases in expenditure next year. Developing leading AI infrastructure will be a core competitive advantage for creating the best AI models and product experiences, and the company will significantly increase investment in 2026 to support this effort. On Google’s side, due to strong and growing demand for its cloud products and services, the company plans to increase capital expenditure by 13% this year to $85 billion, with expectations for further increases in capital expenditure in 2026.

Risk Warning: AI development may fall short of expectations; major firms’ capital expenditures may underperform; industry competition could intensify.