Alibaba Cloud is making significant strides in the artificial intelligence (AI) landscape, unveiling an array of advanced AI models and infrastructure upgrades aimed at bolstering its international market presence. This initiative underscores the company's ongoing commitment to leading AI technology and facilitating widespread digital transformation across various sectors.
Ms. Selina Yuan, the International Business President of Alibaba Cloud Intelligence, emphasized the importance of this launch, stating that the company is presenting a suite of performance updates for its Platform-as-a-Service (PaaS) and AI solutions designed to meet the evolving digital transformation needs of global enterprises. "These upgrades will enhance our ability to deliver superior service efficiency and security, enabling businesses to scale and innovate seamlessly in an AI-driven world. As cloud and AI have become critical factors for global growth, we are dedicated to continuously advancing our core products to meet our customers' changing demands," she remarked.
In its latest announcement, Alibaba Cloud revealed the expansion of access to its foundational models and improved infrastructure for international customers through availability zones in Singapore. This includes the latest models in the Qwen family, which consists of large language models (LLMs) developed by Alibaba Cloud. Key releases include Qwen-Max, a mixture of experts model, QwQ-Plus for logical reasoning, QVQ-Max for image reasoning, and Qwen2.5-Omni-7b, a comprehensive multimodal model. Notably, QwQ-Plus excels in advanced analytical reasoning, adeptly handling complex challenges such as deep analytical question-answering and high-level mathematical problems using highly precise algorithmic solutions.
Meanwhile, QVQ-Max, designed for image reasoning, demonstrates the capacity to tackle intricate multimodal issues efficiently, enhancing reasoning capabilities by accepting image input and producing outputs in a chain-of-thought format. Alibaba Cloud has also significantly upgraded its Platform for AI (PAI) to support Generative AI and LLMs, making AI model usage more accessible and cost-effective.
The PAI-Elastic Algorithm Service (EAS) has introduced distributed inference capabilities with a multi-node architecture to meet the rising demand for super-large models powered by complex mixture of experts structures and long-text processing. This innovation has led to a remarkable 92% improvement in concurrent processing capabilities and a 91% increase in tokens processed per second (TPS) when used with the Qwen2.5-72B model, showcasing a leap in performance and scalability.
Furthermore, the PAI-Model Gallery has been enhanced to offer over 300 modern open-source models, including Qwen and Wan, allowing users to access, deploy, and manage these models smoothly without the need for coding. This initiative includes a variety of processing resources and new features, such as model assessment for performance insights and knowledge transfer from large models to smaller ones, significantly reducing deployment costs.
In a bid to improve data management efficiency, Alibaba Cloud has integrated native AI inference capabilities powered by Qwen into PolarDB, the company’s leading cloud-native relational database. This integration enhances data management performance in the AI era by using machine learning to eliminate inconsistencies between training data and real-world data, which is crucial for inference workflows. The new feature is designed for workloads that focus on text processing, making it suitable for various scenarios, including developing conversational retrieval-augmented generation (RAG) agents, generating text embeddings, and searching for similar meanings.
Additionally, the company has integrated AnalyticDB, its data warehouse, with Model Studio, a platform for Generative AI and application development, serving as a recommended vector database for RAG solutions. This integration directly links the organization’s proprietary knowledge base with existing AI models and tools available on Model Studio, enhancing the efficiency of context-aware application development.
Beyond AI models and infrastructure upgrades, Alibaba Cloud has launched a new suite of AI products in Software-as-a-Service (SaaS) format to accelerate digital transformation across all industry sectors. These AI-powered tools aim to help international customers enhance their data analytics capabilities, automate processes, and efficiently generate content.
Among the new offerings is AI Doc, an intelligent document processing tool that leverages LLMs to analyze various document types, including business reports, forms, product manuals, and procurement documents, accurately extracting the necessary information and rapidly generating reports tailored to specific objectives. For instance, when used alongside Energy Expert, Alibaba Cloud’s sustainable energy solution powered by AI, businesses can quickly process numerous documents and create reports that align with environmental, social, and governance (ESG) reporting frameworks.
Another innovative tool, Smart Studio, is an AI-driven content creation platform that enables applications to seamlessly convert text to images, images to images, and text to videos, thereby enhancing marketing and creativity across various industries, including e-commerce, gaming, education, and entertainment. Examples of applications include leveraging AI to create personalized online shopping experiences for individual customers, virtual fitting of clothing, virtual hair and makeup customization, and easily selecting jewelry, alongside generating images from natural language descriptions or designing products with just a click.
Additionally, Alibaba Cloud has introduced a new AI search function on its official website, utilizing Qwen to assist enterprise customers, particularly small and medium-sized enterprises (SMEs), in quickly discovering solutions and gaining critical insights to inform strategic decision-making. The company is also making accessible and scalable cloud solutions available, with free resources for training in AI and cloud computing.
To foster sustainable growth, Alibaba Cloud has announced an enhanced partner incentive policy aimed at better supporting retailers and distributors by increasing flexibility, enhancing commission rates, and providing more opportunities for rewards to promote mutual growth. This commitment to empowering partners is further supported by training resources and support designed to enhance partner capabilities and drive success within the continuously evolving ecosystem.
In February 2025, Alibaba Cloud announced a significant investment of $5.3 billion (approximately 380 billion baht) over the next three years to develop its cloud computing and AI infrastructure, reaffirming the company's long-term commitment to technological innovation. This substantial investment surpasses the total AI and cloud spending of Alibaba over the past decade, confirming the company’s dedication to advancing AI adoption and strengthening its position as a leading global cloud provider. Currently, Alibaba Cloud operates a global infrastructure network across 87 availability zones in 29 regions worldwide.
