Alibaba CEO Eddie Wu announced the company will train a massive AI model with 5 trillion to 10 trillion parameters, signaling a major expansion into artificial intelligence infrastructure.
Alibaba Group outlined an ambitious artificial intelligence strategy spanning model development, chip manufacturing, and data center expansion.
The planned 5T-10T parameter model would rank among the largest language models in development globally. For context, OpenAI's GPT-4 uses substantially fewer parameters, though exact specifications remain proprietary. Scaling to this magnitude requires massive computational resources and represents a significant capital commitment.
The announcement reflects intensifying competition in AI development. Chinese tech giants including Baidu, Tencent, and Bytedance have all accelerated AI investments. Meanwhile, Western competitors like OpenAI, Google, and Meta continue advancing large language model capabilities.
Wu's roadmap extends beyond models. Alibaba is investing in semiconductor design and data center infrastructure to support training and deployment. This vertical integration approach aims to reduce dependency on external chip suppliers and optimize performance for its specific AI workloads.
The initiative aligns with China's broader AI ambitions. Beijing has designated artificial intelligence as a strategic priority, with recent policy documents emphasizing domestic capability development and technological independence.
Key challenges remain. Training models at this scale demands substantial electricity, cooling systems, and specialized hardware. Supply chain constraints for advanced semiconductors continue affecting tech companies globally. Alibaba's progress will depend on securing adequate resources and navigating evolving regulations around AI development.
The company has previously invested in AI research through its DAMO Academy, focusing on machine learning, computer vision, and natural language processing. This new initiative represents an escalation of those efforts into large-scale foundational model development.
Alibaba did not provide specific timelines for model completion or deployment details. The company may use the model internally across its e-commerce, cloud computing, and logistics operations, or potentially offer it as a commercial product.
Advances in model optimization and open-source tooling have made state-of-the-art AI models accessible on personal devices, eliminating the need for cloud infrastructure or enterprise-grade servers.
Mathematician Terence Tao questions the necessity of human mathematicians in an era of advancing AI and computational tools. The post generated significant discussion on Hacker News, with 87 comments debating automation's impact on mathematical research.
Liberal Democrat leader Ed Davey will propose a global nuclear-style treaty to halt super-intelligent AI development at his party's Brighton conference. He will criticize the government for relying on Donald Trump and tech industry figures to address the issue.
Xiaomi has released two open-weight omnimodal models—MiMo-V2.6 Pro and Flash—with the Pro variant claiming performance parity with Anthropic's Opus 5 and OpenAI's GPT-5.6 Sol on agent benchmarks.