September 23, 2026 — Alibaba Cloud will establish its first cloud regions in Türkiye, Finland and the Netherlands over the next 12 months as the company expands its cloud and artificial intelligence infrastructure outside China.
The company announced the plan on Wednesday at the 2026 Apsara Conference in Hangzhou, China.
Alibaba Cloud will also expand its data-centre footprint in Malaysia, Germany, the United Arab Emirates, France and Hong Kong. The company said the expansion will increase its global computing capacity and strengthen its ability to provide localised cloud and AI services.
Alibaba Cloud currently operates 107 availability zones across 31 regions worldwide, according to the company’s latest announcement.
Alibaba Cloud Chief Technology Officer and President of International Business Feifei Li said the company was expanding its footprint to bring computing resources closer to customers and partners as companies move AI applications from experimentation towards deployment.
The company did not describe the three planned locations as dedicated AI-only regions. Instead, they are new cloud regions intended to increase computing capacity and support both cloud and AI workloads.
South China Morning Post reported that the Netherlands will be the first of the three new locations to open, with the first data centre scheduled for October 2026.
Alibaba Cloud has not publicly identified the exact city, data-centre operator, power capacity or number of availability zones for the planned Türkiye and Finland regions in its September 23 announcement.
The expansion will add to an existing European footprint. Alibaba Cloud currently lists Frankfurt with three availability zones, London with two and Paris with two. Its current infrastructure documentation also lists Dubai with two availability zones and Riyadh as a partner region with two zones.
Alibaba Cloud launched its Paris cloud region in June 2026 with two availability zones. The company said the Paris deployment was intended to serve growing demand from European customers in areas including retail, software development, AI-powered creative platforms and sports technology.
The new announcement therefore adds capacity to an infrastructure network that has already been expanding rapidly across Europe and other international markets.
The United Arab Emirates is also receiving additional capacity, but the announcement does not introduce a new UAE cloud region. Alibaba Cloud already operates a Dubai region.
In October 2025, Reuters reported that Alibaba Cloud opened its second data centre in Dubai, nine years after establishing its first facility in the city.
That expansion formed part of Alibaba’s previously announced plan to invest at least RMB380 billion, about US$53 billion, over three years in AI and cloud infrastructure. Alibaba announced the investment programme in February 2025 and said it exceeded its total AI and cloud spending during the previous decade.
Alibaba’s international expansion has continued through 2026. In June, the company said Alibaba Cloud operated 105 availability zones across 32 regions. Its latest September announcement gives the current figure as 107 availability zones across 31 regions.
Alibaba has also expanded in Malaysia and launched new availability zones in several international markets during 2026. The company previously announced plans for new infrastructure in Brazil, France and the Netherlands, alongside facilities in Mexico, Japan, South Korea, Malaysia and Dubai.
The physical infrastructure programme is being developed alongside a broader push into AI models, chips and enterprise AI services.
On September 22, one day before the cloud-region announcement, Alibaba released its full-stack AI roadmap covering chips, cloud infrastructure, foundation models and AI agents.
Alibaba said its next-generation Qwen 4 model is already in training. It also gave a roadmap for Qwen 4.5 and Qwen 5, with the future model series projected to scale to between 5 trillion and 10 trillion parameters.
The same announcement introduced the Zhenwu V900, an AI training and inference processor designed by Alibaba’s T-Head chip unit.
Alibaba says the V900 delivers three times the performance of its Zhenwu M890 predecessor, includes 216 GB of GPU memory and 1,200 GB/s of inter-chip bandwidth, and supports FP8 and FP4 data precisions.
The company said the V900 is scheduled for mass production and commercial release in the first quarter of 2027.
Alibaba also said its Zhenwu AI chips are already serving more than 650 customers across industries including automotive, finance, large language models, embodied intelligence, energy and manufacturing.
The company has designed an upgraded supernode server around the V900, its ICN Switch, Panmai SmartNIC and Zhenyue SSD controller chip. Alibaba said the system can support a supernode cluster of up to 500,000 cards.
Alibaba’s longer-term target is to have global data-centre capacity operated by Alibaba Cloud exceed 20 gigawatts by 2032. The target was announced by Alibaba CEO Eddie Wu at this year’s Apsara Conference.
The company is also restructuring its cloud platform around what it calls an agentic cloud architecture.
Alibaba describes the architecture through three layers: AI Native Cloud for model training and inference, Agent Native Cloud for enterprise AI-agent deployment and security, and a Context Engine for real-time data and long-term memory.
Its Platform for AI, or PAI, has been updated with inference, caching, sample replay and training optimisations. Alibaba said PAI completed real-world post-training of Qwen models in five days.
The company also announced a new Cloud Parallel File Storage system designed for AI model training. Alibaba said it can deliver throughput measured in the hundreds of terabytes per second and more than 100 million input/output operations per second, while reducing enterprise AI storage costs by up to 69%.
Alibaba’s HPN 8.0 Pro networking architecture is described as providing 100 petabits of bandwidth. The company said a single cluster can support more than 130,000 network ports operating at 800G speeds.
For enterprise AI agents, Alibaba introduced AgentCore, a platform for building, operating and managing AI agents across their lifecycle.
It also introduced an Agent Security Center for security and compliance management, and Agent Context, an enterprise context service that connects documents, business systems, chat records and multimodal data.
Alibaba said Agent Context can reduce token usage by up to 67% in areas including customer service, AI coding and data analytics.
Its upgraded OpenLake platform is designed as a unified multimodal data lakehouse. Alibaba said it can reduce total costs by 38% and query response times by 40% compared with traditional architectures.
Alibaba Cloud also announced three AI products for international customers as part of the September 23 expansion.
Smart Studio is designed to let businesses create dedicated Model-as-a-Service platforms and package, price, meter and bill model APIs under their own brands. Alibaba claims its inference architecture can deliver up to 505% higher throughput than standard open-source frameworks running on the same hardware.
Smart Fusion provides a unified model API that allows multiple models to collaborate on a task. Alibaba says it can reduce token spending by around 50% by selecting models according to real-time requirements and exposing them through a single endpoint.
Smart Video is an AI video agent aimed at short-drama studios and advertisers. Alibaba says it can automatically create videos of up to one hour, supports multiple video models including Wan, and supports more than 20 languages.
The company has also announced new commercial collaborations connected to its international AI push.
Panasonic Digital in Shanghai is using Alibaba Cloud’s Model Studio, Qwen models, DataWorks, MaxCompute and Hologres to build AI agents for manufacturing, supply-chain and corporate functions. Alibaba said the deployment reduced documentation time by 80% and increased contract-review efficiency tenfold through its Smart Compare agent.
Unity China is using Alibaba Cloud’s Model Studio to deploy Qwen and third-party foundation models for Tuanjie AI-Codely, an AI-native engineering platform for game development.
Indonesian logistics company Lion Parcel is using Qwen-VL-Plus through Model Studio to automate finance and document-processing workflows.
Malaysia-based Loomi Entertainment Group has integrated Alibaba’s Wan and HappyHorse visual-generation models into its cinematic AI engine, Imaginary. Alibaba said the companies have held two AI-generated-content hackathons and trained more than 100 local content creators.
SHAKE, which provides agentic resource-planning services for small and medium-sized businesses, is using Alibaba Cloud compute and Qwen models to build a proprietary AI model. Its services will also be offered through Alibaba Cloud Marketplace.
Alibaba Cloud separately announced a strategic agreement with AnyMind Group to integrate Qwen into AnyLive, the company’s AI-powered live-commerce platform. Qwen will be one of the models used in AnyLive’s multi-model architecture.
Alibaba’s cloud expansion is taking place as European governments are also increasing attention on data-centre capacity, cloud infrastructure and technology sovereignty.
On June 3, 2026, the European Commission proposed the Cloud and AI Development Act. The proposal aims to strengthen Europe’s cloud and AI ecosystem, accelerate data-centre deployment and introduce an EU-wide framework for assessing cloud and AI sovereignty.
The Commission says Europe needs more cloud and data-centre capacity as AI demand increases. Its proposal aims to at least triple EU data-centre capacity within five to seven years and meet the needs of EU businesses and public administrations by 2035.
The Commission has also identified reliance on non-EU cloud providers as a risk to Europe’s digital autonomy and resilience. The proposal does not specifically target Alibaba Cloud.
For Alibaba Cloud, the European expansion also raises questions about where data is processed and which services will be available locally.
A new cloud region does not by itself establish that all customer data will remain inside Europe. Alibaba’s privacy documentation describes circumstances in which personal data can be transferred outside the European Economic Area or the United Kingdom, subject to applicable safeguards.
Alibaba’s current Model Studio documentation also shows that AI model availability can differ between regions. The existence of a new cloud region therefore does not automatically mean that every Qwen or AI service will be available there at launch.
The expansion comes as Alibaba’s cloud business reports stronger AI-driven demand. In the quarter ended June 30, 2026, Alibaba reported RMB48.437 billion in revenue from AI Cloud and Compute Services, up 45% year over year.
AI-related product revenue reached RMB12.376 billion, representing the twelfth consecutive quarter of triple-digit year-over-year growth for that category.
Adjusted EBITA for AI Cloud and Compute Services increased 133% year over year to RMB5.628 billion.
Those results provide the commercial backdrop to Alibaba’s continued infrastructure spending as the company expands its capacity for AI training, inference, cloud services and agent-based applications.
The September 23 announcement does not provide the individual investment amount, power capacity, physical site or availability-zone count for each of the three new cloud regions. Those details remain to be announced.
What Alibaba has confirmed is the geographic direction: new cloud regions in Türkiye, Finland and the Netherlands, additional capacity in Malaysia, Germany, the UAE, France and Hong Kong, and a broader infrastructure strategy aimed at supporting the company’s expansion across cloud computing, AI models, proprietary chips and enterprise AI applications.
Discover more from Aree Blog
Subscribe now to keep reading and get access to the full archive.
