New-generation communication networks encompass mobile networks such as 5G, 5G-A, and 6G, as well as 10-gigabit optical networks and satellite internet, enabling deep integration of functions including communication, sensing, computing, and intelligence. China's new-generation communication network construction is steadily advancing, with coverage continuously expanding. These networks have been deployed in scenarios such as forest fire monitoring, mountain road management, and full-area maritime connectivity, with integrated applications continuously deepening. More than 26,000 industrial 5G private networks and 1,260 5G factories have been built, with the industrial internet covering all 41 major industrial categories and achieving remarkable results in fields such as new energy vehicle manufacturing and open-pit mining. Meanwhile, 6G technology is accelerating its transition from the laboratory to real-world application, and technologies such as multi-base-station cooperative sensing will in the future provide faster response guarantees for fields such as autonomous driving.

As reported on November 6, the European Central Bank is considering a review of bank loans to data centers to assess the risks arising from shifts in AI market sentiment and valuation fluctuations. Senior officials have discussed the plan internally and with supervised banks. The review may include on-site inspections, focusing on how banks assess future cash flows, collateral values, and potential default risks. The scope of the review may be smaller than the leveraged finance review conducted several years ago, and the specific plan has not yet been finalized.

The Administrative Committee of the Beijing Economic-Technological Development Area has issued measures to accelerate the development of an open-source and open innovation hub for the information technology application innovation industry. These measures propose supporting the construction of platform bases for key information infrastructure industries and establishing common technology empowerment centers in frontier fields. For relevant platforms operating on a substantive basis, support of up to 30 million yuan per year for no more than three consecutive years will be provided, based on various factors.

OpenAI Chief Financial Officer Sarah Friar said on Friday that the company will continue to advance its U.S. data center expansion plans, despite concerns that the Democratic midterm elections could trigger stricter scrutiny of AI infrastructure. On Thursday, approximately $2.5 trillion was wiped off U.S. stock markets amid fears that a Democratic sweep in the midterm elections could slow construction of key AI facilities. Investors are assessing the impact on the chip, cloud, and networking sectors—the biggest beneficiaries of the spending boom. However, signs indicate that despite growing opposition, tech giants remain committed to data center expansion. Google announced on Thursday that it will invest $17 billion to build its first data center in Wyoming.

Google has acquired Crusoe's data center site in Wyoming.

Microsoft is gradually terminating its arrangement to rent AI computing power from Oracle—a move that is part of a broader reassessment of its cloud partnerships. The Oracle computing power previously rented used older-generation Nvidia chips and no longer meets Microsoft's specifications. Having committed more than $80 billion to rent computing power from other cloud companies, Microsoft is now rethinking its related strategy, dissatisfied with the high costs and delivery delays of new cloud suppliers.

On November 7, foreign media reported, citing people familiar with the matter, that Microsoft has stopped renewing its AI compute leasing contract with Oracle and has been gradually decommissioning the leased computing resources over the past few months. According to these sources, the servers provided by Oracle use an older generation of Nvidia chips and no longer meet Microsoft's current technical specifications. The two parties reached an agreement in 2023 to supply computing power for the new version of the AI search engine Bing, though the contract value was not disclosed. Oracle responded that demand for its AI infrastructure remains strong and that older-generation GPUs still have a market. Microsoft is reassessing its external AI compute partnerships, having previously committed more than $80 billion in compute leasing spending to emerging cloud service providers such as CoreWeave, Nscale, and Nebius. Earlier reports indicated that Microsoft management has grown increasingly dissatisfied with the high costs and project delays of some suppliers, has recently tightened reviews of data center projects, exited certain high-risk projects, and is reconsidering whether to sign new compute cooperation agreements. Meanwhile, Oracle's cloud business is expanding rapidly, but heavy capital expenditures have pushed free cash flow into negative territory, intensifying investors' concerns about the returns on its AI infrastructure investment; its stock price has fallen by more than half from this year's high. Following the news, Oracle's stock price quickly turned negative, with intraday losses once widening to 4.6%. It is currently down about 3%, having previously risen about 3.5%.

The massive demand for chips and storage resources from AI data centers is reshaping the competitive landscape of the global semiconductor industry. AMD, once known for PC processors and gaming graphics cards, has now officially pivoted to artificial intelligence and data center businesses. CEO Lisa Su explicitly stated that the company is now a "data center first" enterprise.

On November 6, approximately 100 environmental and social activists breached the construction site of Microsoft's new data center in Amsterdam, the Netherlands, with some scaling cranes to unfurl banners protesting artificial intelligence. The incident underscores growing opposition within European society to the construction of large-scale data centers and the expansion of the AI industry.

Meta executives have cautioned that the outcome of the 2026 U.S. midterm elections could make it more difficult for tech companies to build data centers. Following the Democratic Party's victory in these midterms, disputes over data centers and AI infrastructure have intensified, and Meta is concerned that political and social opposition could impact its future investments.

Jiechuang Intelligent announced on November 6 that it has entered into a "Computing Power Technical Service Contract" with Company H1 for the provision of computing power services. The provisional total contract value is RMB 4.915 billion (tax inclusive), with a term of 60 months. Each batch of orders covers a service period of 60 months, with monthly settlement. This is a routine business contract that does not require approval by the board of directors or the shareholders' meeting, nor does it constitute a related-party transaction or a major asset restructuring. Due to trade secrets, information such as the name of Company H1 has been exempted from disclosure.

Tonghe Technology has announced plans to invest 1.0025 billion yuan in the Shijiazhuang High-tech Industrial Development Zone to build an industrialization project for integrated computing power and energy storage intelligent complete power supply equipment. The project will include the research, development, and production of data center power supply and distribution systems and modules (Phase II), the industrialization of new energy storage systems, and the industrialization of intelligent integrated power supply systems.

Tongyu Communication announced on November 6 that it plans to invest 500 million yuan of its own funds to establish a wholly-owned subsidiary, Shanghai Yuyuan Zhisuan Technology Co., Ltd. (the name is subject to registration with the market supervision authority). Yuyuan Zhisuan will serve as the company's computing power business platform, pursuing business exploration and industrial deployment in the intelligent computing services sector. The company will align with its development plan, leverage industrial synergy advantages, strengthen communication and collaboration with relevant partners, and promote steady business growth.

AMD CEO Lisa Su says the company has transformed into a "data center-first" enterprise and expects its data center business to grow 50% quarter-over-quarter in the fourth quarter of 2026 and more than double in 2027.

NVIDIA is exploring a way to rapidly deploy chip computing power by repurposing old telecom buildings, and has approached private equity firm Apollo Global Management. The plan is to leverage more than 130 telecom buildings owned by Brightspeed—a near-bankrupt broadband operator under Apollo—which already have wiring in place and ample power supply, to build a small AI data center network covering 20 U.S. states. This would allow the company to sidestep the cumbersome approval and regulatory processes for new data center construction that are about to be introduced in many parts of the United States. NVIDIA's assessment concludes that these buildings collectively can provide approximately 1 gigawatt of available power—enough to support Vera Rubin chip computing clusters worth tens of billions of dollars—and it plans to install server racks in some of the buildings as early as the first quarter of 2027. The plan may also incorporate technology from Groq, an AI chip startup from which NVIDIA obtained licensing at the end of last year. The project is positioned as an inference network, with computing power mainly used to run already-mature AI models. Potential customers include cloud providers and AI research and development companies urgently seeking NVIDIA computing power, such as SpaceX, Anthropic, and OpenAI.

On November 6, Anoky announced plans to establish a wholly-owned subsidiary, Yantai Anoky Intelligent Computing Technology Co., Ltd. (tentative name), which will serve as the implementing entity for an intelligent computing center project with a total investment of 580 million yuan. The company will fund the project through its own funds or self-raised capital, investing in installments according to construction progress. This investment does not constitute a related-party transaction or a major asset restructuring.

Scattered monitoring points, challenging public network coverage, and year-round unattended operations are common hurdles in remote and harsh industrial environments. Dayu Semiconductor's FishHOC narrowband self-organizing network solution leverages site relay to wirelessly transmit data from remote areas back to the control center.

NVIDIA has submitted a written proposal to U.S. broadband operator Brightspeed to convert some aging telecom central offices into small AI data centers, forming an inference network covering up to 20 U.S. states. People familiar with the matter said NVIDIA estimates these facilities have a combined available power of about 1 gigawatt.

The openssh-server package on Alpine Linux contains a password authentication bypass vulnerability in builds where PAM is not enabled. To exploit it, an attacker must meet three conditions: PasswordAuthentication is enabled, an account exists whose password field is set to *, and at least one account on the local machine uses a $5$ or $6$ hash. Once these conditions are satisfied, the attacker can log in without the correct password by submitting an overly long passphrase. Affected versions include Alpine 3.16 through 3.24 and the edge version. Systems running openssh-server-pam with UsePAM set to yes, as well as systems where password login has been disabled, are not affected. No CVE has yet been assigned to this vulnerability. The root cause is that, without PAM, the salt of another account is borrowed, and musl's crypt() returns * for overly long passphrases. A related patch has been released, and users should upgrade the package and restart sshd. Those unable to upgrade immediately can disable password authentication as a workaround. Notably, the alpine user in the official tiny-cloud image carries the risk of being exploited to escalate privileges to root.

On November 6, Yonhap News Agency reported that South Korea's Deputy Prime Minister and Minister of Science and ICT, Bae Kyung-hoon, stated that amid growing concerns over a global slowdown in AI investment, South Korea should not delay investment in AI data centers. Bae emphasized that South Korea must neither halt nor postpone investment under the influence of the current atmosphere. His remarks came in response to recent concerns that the aftermath of the U.S. midterm elections could lead to tighter regulation of AI data centers and declines in related corporate stock prices, triggering worries about a contraction in AI infrastructure investment. Bae pointed out that even if the United States hesitates, countries such as France and Canada may instead see it as an opportunity, and global competition in AI investment is unlikely to cool easily. He assessed that market expectations for AI remain high, and minor unsettling factors could cause severe market fluctuations, but the long-term trend will be upward.