Correcting and Replacing: Huawei Cloud Introduces Token Service in Asia Pacific

Following the rollout of AI Compute Service in Hong Kong (China), Thailand, Indonesia, and Singapore in 2025, Huawei Cloud MaaS (Model-as-a-Service) is now available, leveraging Huawei’s in-house acceleration engine to support mainstream models and provide customers with stable and high-quality token services.

Huawei Cloud continues to strengthen its collaboration with high-quality open-source models in the industry. Six models from three categories—GLM, DeepSeek, and Qwen—are the first to be supported by MaaS. These models are designed for two major application scenarios: intelligent Q&A and AI coding. They offer one-click access and out-of-the-box model hosting services, covering the entire lifecycle of model management from deployment to inference, fine-tuning, and evaluation. With elastic compute and pay-per-use billing, enterprises and developers can quickly integrate AI capabilities into their service systems without having to worry about the underlying infrastructure. This makes AI accessible to everyone.

Notably, this release includes support for the latest GLM-5 model. GLM-5 has attained state-of-the-art (SOTA) performance in coding and agent capabilities, making it an ideal general-purpose agent foundation for complex system engineering and long-text agent tasks in enterprise environments.

Huawei Cloud’s long-term strategic investment in AI has resulted in a comprehensive suite of AI solutions, including CloudMatrix AI Infra, MaaS, ModelArts (a model training and inference platform), CodeArts (a coding agent), AgentArts (an agent platform), and DataArts (a data agent).

CodeArts and AgentArts will be launched outside the Chinese mainland in the second half of 2026.

Huawei Cloud has established an optimized cloud infrastructure network in the Asia Pacific region. This network comprises five Regions and 18 availability zones (AZs) located in Singapore, Thailand, Hong Kong (China), Indonesia, and the Philippines, ensuring a 50 ms access latency. Additionally, in Malaysia, Huawei Cloud and partners have jointly built a local cloud.

Huawei Cloud AI Compute Service have helped many Asia Pacific customers elevate their cloud journey from simply migrating to the cloud to using AI well. They have benefited from improved productivity and better adaptation to changing environments.

iFLYTEK, a leading intelligent voice and AI company, selected Singapore as the first destination for its global expansion strategy in 2023, naturally extending its collaboration with Huawei Cloud from China to a global scale. Huawei Cloud AI Compute Service helped iFLYTEK quickly deploy their large language model training resources within two weeks. The training remained stable for 60 days without interruption, ensuring zero faults during the release of major versions.

Green and Smart Mobility (GSM), an emerging green mobility platform, chose Huawei Cloud to reduce costs, enhance efficiency, and achieve multi-cloud flexibility. GSM has developed a comprehensive fleet management platform using AI and IoT. By utilizing Huawei Cloud IoTDA and ModelArts AI technologies, GSM analyzes violations and generates risk alerts to ensure driver and passenger safety, thereby fulfilling its mission of promoting green and safe transportation.

Huawei Cloud will continue to invest in AI innovation and provide a secure, stable, and intelligent cloud foundation to help enterprises in Asia Pacific achieve sustainable business growth in the AI era.

Hashtag: #HuaweiCloud #AI #TokenService

The issuer is solely responsible for the content of this announcement.