Alibaba is outlining plans for next-generation artificial intelligence infrastructure, including an AI model expected to reach up to 10 trillion parameters, and the introduction of new chip technology to support larger-scale computation. It also signals additional capacity in its broader data-centre buildout.
The reports present the initiative as part of a wider expansion of Alibaba’s AI stack, pairing model scaling with dedicated hardware. One outlet frames the model size range as 5 trillion to 10 trillion parameters, while another describes it as up to 10 trillion parameters. Both emphasize that the company is investing in both compute and infrastructure, aiming to enable training and deployment at higher performance levels.
While the sources differ slightly in how they specify the parameter range, they align on the core elements: a move toward much larger AI models, unveiling new chips, and expanding data-centre capacity as supporting infrastructure. Neither source provides additional independent verification of performance claims or a detailed deployment timeline.
