- Alibaba Cloud is adding data centres in Malaysia, the Philippines, and an AI hub in Singapore.
- It plans to train 100,000 AI workers a year and invest $53B in AI.
Alibaba Cloud is expanding its presence in Southeast Asia with new data centres in Malaysia and the Philippines, part of a broader push to meet rising demand for AI services in the region.
The company recently launched its third data centre in Malaysia and plans to open a second in the Philippines by October. These follow similar investments in Thailand, Mexico, and South Korea earlier this year.
Alibaba has committed to spending US$53 billion on AI infrastructure over the next three years. It’s also positioning Southeast Asia as a key market, even after shutting down data centres in Sydney and Mumbai last year, as reported by South China Morning Post.
AI hub and talent development in Singapore
As part of its 10-year anniversary in Singapore, Alibaba Cloud announced a new AI Global Competency Center (AIGCC) in the country. The centre is intended to support over 5,000 businesses and 100,000 developers, offering tools for building and deploying AI systems.
It also includes an AI Innovation Lab that will provide curated datasets, usage credits, and support services. The AIGCC is expected to work with more than 1,000 companies and startups, with plans to introduce over 10 AI agents for industries like healthcare, logistics, and finance.
To build a larger talent pool, Alibaba Cloud says it will partner with over 120 institutions to train 100,000 AI professionals annually.
New AI tools and system upgrades
Alibaba Cloud also rolled out upgrades to its AI infrastructure and software tools.
Its real-time data service, Data Transmission Service (DTS), now supports an “One Channel for AI” feature that can convert various types of data—text, images, audio, video—into formats that can be used for AI training and Retrieval-Augmented Generation (RAG) applications. The goal is to reduce technical complexity and speed up deployment.
On the AI inference side, Alibaba’s Platform for AI (PAI) introduced updates to support large models and complex architectures like Mixture of Experts. A new Model Weights Service is also available to cut cold start times and improve scaling, with test results showing up to 90% faster performance in some cases.
The company’s ninth-generation Intel-based Elastic Compute Service (ECS) instance is also expanding to more markets, including Japan, the UAE, and the UK. Since its April launch, nearly 10,000 businesses have adopted the new instance, which offers better computing efficiency and faster networking for AI, HPC, and database workloads.
Green AI research and challenges
A new study released during Alibaba Cloud’s global summit looked at how businesses are approaching “green AI”—AI systems designed to reduce environmental impact. Conducted by Forrester and commissioned by Alibaba, the study surveyed more than 460 IT and business leaders globally.
While most respondents agreed that green AI is important, many said they’re still in early stages. Some of the top challenges include sourcing sustainable hardware (80%) and improving data centre energy efficiency (73%).
Capability gaps were also common. Around three-quarters of respondents said their organisations lack the knowledge or skills to build and operate green AI systems. The study recommended several steps to close the gap, including using renewable energy in data centres, building smaller models, and improving collaboration on standards and open-source tools.
AI in practice: Customer examples
Alibaba Cloud’s AI offerings are being used by a growing number of customers across Asia and beyond.
Indonesia’s GoTo Group migrated its core business intelligence data platform to Alibaba Cloud’s MaxCompute. The company says the migration, which moved tens of petabytes of data over six months, helped improve cost efficiency and system performance with no downtime. GoTo Financial has also moved its lending systems to Alibaba Cloud, using PolarDB and Tair to support over 500 microservices with low latency.
VisionTech, a generative AI startup in Singapore, uses Alibaba Cloud infrastructure to scale its multilingual AI bots across Southeast Asia. The company says it cut infrastructure costs by more than 25% and now uses Alibaba’s Qwen model to manage real-time translation across English, Chinese, Malay, and Japanese.
Japanese tech provider FLUX is also working with Alibaba Cloud to bring the Qwen model to local businesses. FLUX plans to build its own LLM product using Alibaba’s tools and apply it to core operations for clients across industries.
In the Middle East, Alibaba Cloud signed an agreement with Al-Futtaim, a diversified business group based in Dubai, to support AI development across its business units. The deal includes access to Alibaba’s cloud and AI infrastructure, as well as open-source frameworks and training support.
