Huaqin Technology revealed on an interactive platform that, by harnessing its comprehensive full-stack strengths and pioneering R&D prowess, the company has positioned itself as one of the select few suppliers in the industry capable of independently devising both computing nodes and network nodes. This unique capability grants Huaqin Technology a significant head start in the super-node product domain. Presently, the company's super-node products are primarily catering to the needs of leading domestic cloud service providers. Initial product shipments commenced in the second quarter of this year, followed by a remarkable upswing in growth starting in the third quarter, and a sustained scaling up is anticipated in the fourth quarter.

On September 23, it was reported that during the 2026 Ecological Partner Conference hosted by Jishi Media, a strategic agreement was forged among Kunlun Yuan, Jingjia Micro, Jishi Media, and Vision Power. The consensus focused on the construction of computing power infrastructure and system integration for the China Radio and Television (Jilin) Cultural Media Intelligent Computing Platform project. This agreement delineated the project's initial scope, paving the way for the formal signing of agreements and the commencement of project construction.

As per the project blueprint, the initial phase will involve the local deployment of 100P of intelligent computing capacity, complemented by access to an additional 700P of external computing resources. This collaborative effort aims to supply computing power nodes for radio and television, facilitate intelligent production within the converged media landscape, and support the creation of AIGC (Artificial Intelligence Generated Content) micro-short dramas. The media computing power hub will be developed in alignment with China Radio and Television's '1+7+31+N' computing power system framework. It will be integrated into the national unified layout as a pivotal node for radio and television computing power scheduling, offering comprehensive services that combine 'media-specific computing power' with 'intelligent content computing power' for the industry. Furthermore, it will foster intensive computing power operations and cross-regional cooperation.

This collaborative endeavor not only clarifies the construction trajectory and initial computing capacity scale but also establishes a solid foundation for the project's successful implementation.

Tencent Security has recently revealed that the Tencent WorkBuddy Enterprise Edition has made significant strides in its transformation into a full-fledged office suite. Notably, Tencent Docs, Lexiang, Cloud Drive, the Security and Efficiency Operation Platform, and the AI-powered design assistant Ardot have been seamlessly integrated as part of this upgrade. Following the upgrade, the pricing structure for the WorkBuddy Enterprise Edition remains steady, while offering a unified enterprise office system that encompasses documents, knowledge base, cloud storage, design tools, security features, and efficiency-enhancing functionalities.

AI company Anthropic is in talks with Stream Data Centers, planning to lease data centers with a capacity of up to 1 gigawatt to solidify control over more data centers. Additionally, Anthropic has signed a data center lease agreement in Australia, with a planned capacity of 2.16 gigawatts.

Media sources indicate that Anthropic is currently engaged in negotiations aimed at solidifying its dominance over an expanded network of data centers. The company has already entered into several preliminary lease agreements for data centers across the United States, boasting a combined capacity that surpasses 1 gigawatt. Furthermore, it is actively engaged in discussions regarding leasing arrangements with various entities, including Stream Data Centers. In a strategic move, Anthropic has also secured a lease for a massive 2.16-gigawatt data center campus in Australia. To fuel its ambitious gigawatt-scale data center expansion plans, the company is seeking financial backing from Google.

On September 23, reports emerged that Alibaba is intensifying its data center expansion efforts in Europe and the Middle East, as part of a broader initiative to advance its global infrastructure development, with a target capacity of 20 gigawatts. During the Cloud Town Conference held in Hangzhou on Wednesday, Alibaba executive Li Feifei revealed that over the next 12 months, Alibaba plans to establish its inaugural batch of cloud regions in Turkey, Finland, and the Netherlands. Additionally, the company will broaden its data center footprint in Malaysia, Germany, the United Arab Emirates, France, and Hong Kong, China. This strategic maneuver is designed to fortify localized cloud and AI service capabilities.

Presently, Alibaba's global cloud network already spans a diverse array of regions, stretching from Japan across Southeast Asia to the Middle East. In recent times, the company has inaugurated new facilities in Brazil and France, specifically tailored to deliver generative AI services to local enterprises. Li Feifei underscored the company's commitment, stating, "We are dedicated to collaborating closely with the global ecosystem to make AI more scalable, practical, and accessible for businesses worldwide."

At the Apsara Conference, Alibaba Cloud proudly presented its enterprise-grade Agent platform, AgentCore. This innovative platform is designed to significantly boost business productivity by fostering seamless human-machine collaboration and offering a comprehensive suite of tools for Agent development and governance.

Alibaba will establish new data centers in Finland and the Netherlands, and expand its data center Layout (layout) in Germany, France, and the UAE to meet global AI and cloud computing demands.

According to the report, intelligent agents are revolutionizing the computing power infrastructure, propelling a substantial surge in Token consumption and, consequently, significantly accelerating the growth rate of China's intelligent computing capacity. Computing power has now emerged as a national strategic asset, ranking on par with essential infrastructures such as water conservancy systems, power grids, and railways. The report forecasts that with the swift proliferation of intelligent agents and the corresponding rise in Token consumption, the demand for computing power will persistently escalate. It is anticipated that China's intelligent computing capacity will soar to 2,576.5 EFLOPS by 2026, marking an 87.9% year-on-year increase. By 2030, this figure is expected to further skyrocket to 10,524.3 EFLOPS. As a result, the compound annual growth rate of China's intelligent computing capacity from 2025 to 2030 is projected to climb from the previously estimated 46.2% to 50.3%.

The UK's new cloud service provider, Nscale, is set to go public, and its heavy reliance on two major clients for revenue will face scrutiny from public investors. Nscale was spun off from Australian cryptocurrency mining firm Arkon Energy two years ago and operates independently. According to documents filed with the U.S. Securities and Exchange Commission, Nscale has secured contracts worth over $103 billion in total. However, approximately 85% of these orders are concentrated in two major deals: one is a $43.8 billion contract to supply computing power services to Microsoft by 2033, and the other is a $44.6 billion computing power supply agreement with Anthropic.

The '15th Five-Year Plan' has distinctly charted the course for the evolution of next-generation communication networks. As a pivotal element among the nation's 'six key networks,' the next-generation communication network acts as a linchpin, linking these networks together and serving as the fundamental backbone for Digital China and the ubiquitous Internet of Everything ecosystem. The plan aims to fully realize a robust and high-performance next-generation communication network by 2030, setting forth 13 crucial benchmarks, among which is the ambition to attain an information and communications industry revenue of RMB 4.1 trillion.

Starcloud, a pioneering space computing startup based in the U.S., is forging ahead with an audacious plan: to deploy Bitcoin mining equipment in space and strive to become the inaugural company worldwide to carry out Bitcoin mining in orbit. Backed by prominent institutions like NVIDIA, Starcloud's long-term vision extends beyond merely launching mining rigs into space. The company aspires to capitalize on the abundant solar energy and the vacuum conditions of space to construct expansive orbital data centers. This strategic move would enable energy-intensive operations, such as artificial intelligence computing and cryptocurrency mining, to be conducted efficiently in the space environment.

Cloudflare has officially launched its Python Workers service, with Python now becoming a core language for its developer platform. It seamlessly integrates with the platform's full range of services and directly supports running mainstream Python Web frameworks such as FastAPI, Django, and Flask. Based on Cloudflare's existing WebAssembly technology and integrating the Python interpreter compiled by Pyodide Wasm, developers do not need to adjust their development approach or refactor code for the edge platform. The official version released this time optimizes the development experience by eliminating the need for manually written type conversion code. The relevant logic has been encapsulated in the runtime environment and Python SDK, allowing developers to directly call platform services without interacting with JavaScript. The service is compatible with WSGI and ASGI standards in the Python ecosystem through a lightweight bridging layer, with Cloudflare's platform handling Web server functions uniformly, including load balancing and automatic scaling. Addressing the lack of TCP sockets due to WebAssembly sandbox environment limitations, Cloudflare has achieved a breakthrough through Hyperdrive technology, enabling Python Workers to run database drivers such as asyncpg and aiomysql smoothly, providing full support for relational database operations. Developers can directly use familiar Python frameworks to develop APIs and integrate with platform services, without needing to configure server infrastructure themselves throughout the process.

At the 2026 Apsara Conference, Aliyun introduced CPFS, a cutting-edge, high-performance storage solution tailored for next-generation AI training clusters. This storage system boasts throughput in the hundreds of terabytes per second (TB/s) and tens of millions of input/output operations per second (IOPS), with a single file system capable of scaling up to 100 petabytes (PiB). It significantly accelerates the training process for models that utilize millions of cards, slashing the average model startup time by half, boosting peak computing power utilization by 30%, reducing AI storage costs by up to 69%, and doubling the capacity to handle business operations.

CPFS leverages a fully in-house developed architecture that segregates data and metadata services, allowing a single file system to manage trillions of files efficiently. It facilitates unified management of training data, seamless scaling in sync with GPU clusters, as well as intelligent lifecycle management and tiered cold-hot storage functionalities.

Additionally, Aliyun unveiled KVCacheStore, a KVCache acceleration engine designed for large-scale model inference. Functioning as a novel G3.5 storage layer, it adeptly handles massive caches, delivering a throughput of 40GB/s per compute node and achieving millions of queries per second (QPS). A single instance of KVCacheStore can support KV storage on a scale of hundreds of billions, with measured cache hit rates showing an improvement of over 20%.

Aliyun is committed to fostering collaborative innovation across computing, networking, and storage domains to deliver high-performance, scalable, and cost-effective storage solutions for AI training and applications.

On September 22, during the 2026 Apsara Conference, Alibaba Cloud proudly introduced its latest innovation—a new-generation, high-performance storage solution known as CPFS (Cloud Parallel File Storage). This cutting-edge system is meticulously crafted to cater to the demands of next-generation AI training clusters. With an impressive throughput capability reaching up to hundreds of terabytes per second (TB/s), tens of millions of IOPS (Input/Output Operations Per Second), and an expanded single file system capacity of up to 100 petabytes (PiB)—five times larger than its predecessor—CPFS sets a new benchmark in storage performance.

It significantly accelerates the training process for models comprising millions of cards, while also providing robust support for large-scale model and multimodal data processing tasks. In real-world applications, CPFS has demonstrated remarkable efficiency gains, including an average reduction of 50% in model startup times, a 30% boost in peak computing power utilization, a substantial 69% decrease in AI storage costs, and a remarkable doubling of business carrying capacity.

On September 22, Accenture announced its collaboration with Amazon Web Services (AWS) to provide AI and cloud transformation services for mid-sized enterprises with annual revenues ranging from $300 million to $3 billion through the newly established Accenture Edge business. The collaboration introduces six directly deployable solutions, including data discovery, virtual machine migration, customer experience optimization, cloud cost optimization, AI security testing, and cloud security infrastructure.

On September 22, Xianhe Environmental Protection made an announcement stating that its fully-owned subsidiary, Xingtai Zhisuan, has set its sights on acquiring 63 GPU servers. The acquisition will be facilitated through a third-party financial leasing arrangement from Hebei Bingji, a related entity. Each server is priced at 2.66 million yuan, culminating in a total investment of 168 million yuan. The primary objective behind this procurement is to yield investment returns by offering computing power services. This transaction has already received the green light from the board of directors and now awaits the shareholders' meeting's scrutiny for final approval.

On September 22, Oracle Health Sciences and Wipro announced a partnership to strengthen Wipro's pharmacovigilance service capabilities through Oracle Health Sciences Argus Business Process Services (BPS), providing global life sciences companies with drug safety data processing, regulatory reporting, and compliance support. The two parties will adopt a pre-configured cloud-based Argus BPS environment, completing deployment before customer onboarding to shorten project implementation cycles. Wipro's service scope includes case processing, report generation, regulatory submissions, pharmacovigilance lead support, and post-marketing surveillance.

On September 22, reports emerged indicating that Xingyun Technology had earlier entered into a master framework procurement agreement with a state-owned supplier. This agreement covered the purchase of 128 high-end computing power servers along with their accompanying services. More recently, Xingyun Technology has successfully signed an additional 3 contracts, acquiring 43 sets of GPU computing power services. Each of these services comes with a 5-year service term. The cumulative tax-inclusive value of these three contracts amounts to an impressive 453 million yuan.

On September 22, Coherent made a significant announcement regarding the launch of its PhotonLink™ integrated optical platform. This cutting-edge platform seamlessly integrates the company's core photonic technologies, precision assembly processes, and rigorous testing capabilities, specifically designed to offer comprehensive optical solutions tailored for next-generation AI data centers. It supports a range of advanced optical configurations, including co-packaged optics (CPO), near-packaged optics (NPO), and optical interconnects between chips. Revenue contributions from this platform are projected to commence in the fourth quarter of 2026.

It has been reported that Coherent's data center manufacturing base has achieved a remarkable milestone, having shipped over 300 million indium phosphide lasers and more than 1 billion indium phosphide and gallium arsenide photodetectors. Currently, the company is actively expanding its production capacities for 6-inch indium phosphide wafers and specialty fibers, in a strategic move to bolster the production ramp-up of the PhotonLink platform.