Alibaba Cloud announced significant expansions to its Southeast Asian infrastructure, including the opening of its third data center in Malaysia on July 1 and plans for a second facility in the Philippines slated for October 2025. These developments respond to rising regional demand for cloud and AI services across Southeast Asia.
The new data centers were among several key announcements at Alibaba Cloud’s Global Summit held recently in Singapore, marking a decade of the company’s presence in the city-state. Singapore also hosts Alibaba Cloud’s international headquarters, reinforcing its role as a strategic hub for regional digital transformation.
With these expansions, Alibaba Cloud will increase its global network to 90 availability zones across 29 regions, further cementing its leadership in the global cloud market.
Alibaba Cloud also announced the launch of its first AI Global Competency Center (AIGCC) in Singapore. The AIGCC is designed to support over 5,000 businesses and 100,000 developers, accelerating AI adoption across enterprises of all sizes and addressing the growing global demand for AI talent.
“Over the past decade, Singapore has been both an innovation center and a gateway to the region’s digital economy. As we celebrate this important milestone, we reaffirm our commitment to empowering businesses of all sizes and verticals while advancing cutting-edge AI innovations and driving sustainable digital transformation in Singapore for years to come,” said Selina Yuan, President of International Business at Alibaba Cloud Intelligence.
Alibaba Cloud will further invest over $60 million to strengthen its partner network and foster AI innovation during the current fiscal year.
Enhanced AI Infrastructure and Solutions
Alibaba Cloud also introduced significant enhancements to its cloud services, including improvements to Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) products, as well as advanced AI platform capabilities, which will simplify the business adoption of AI technologies.
Its Data Transmission Service (DTS) now features a pioneering “One Channel For AI,” streamlining data preparation by automating pipeline creation for various data types, converting them into vector databases suitable for AI applications.
In parallel, Alibaba Cloud’s Platform for AI (PAI) has upgraded its Elastic Algorithm Service (EAS) inference capabilities, enhancing support for complex models, such as Mixture of Experts (MoE). The new Expert Parallel optimization, combined with a prefill-decode disaggregation framework, notably improved the efficiency of large language models. Tests demonstrated performance of over 15,000 tokens per second with Qwen3 235B while maintaining sub-50 millisecond latency per token.
Additionally, PAI-EAS introduced a Model Weights Service, greatly improving deployment efficiency. In tests with Qwen3-32B, it achieved a 91.4% faster cold start and near-instant scaling.
Alibaba Cloud’s 9th Generation Intel-based Enterprise Elastic Compute Service (ECS) instances are expanding globally from July, now covering eight additional markets in Asia, Europe, and the Middle East. The latest ECS generation offers 20% greater computing efficiency and up to 50% improved performance for HPC workloads, search recommendations, and Redis databases.
Advancing Sustainability with AI
Alibaba Cloud introduced an AI-powered ESG Reporting solution within its Energy Expert platform. This new tool, leveraging Alibaba’s flagship large language model, Qwen, helps organizations streamline ESG reporting and align with international standards such as ISSB, GRI, and SASB.
The summit also featured the release of a global study on Green AI, conducted by Forrester Consulting in collaboration with Alibaba and the Alibaba-NTU Global e-Sustainability CorpLab (ANGEL). The survey of over 450 business and IT leaders highlighted a growing awareness of AI’s environmental impact and sustainability potential, although many organizations face significant execution challenges. While 84% recognized the importance of Green AI, 69% remain in the initial adoption stage of AI, citing barriers such as a lack of sustainably sourced hardware (80%) and the difficulty of optimizing data center energy usage (73%).
