Alibaba is making a bigger push into artificial intelligence, and this time the company is focusing on more than just AI software. The Chinese technology giant has announced a new AI chip and revealed plans for a much larger model as it works to build more of the technology needed to run advanced artificial intelligence systems itself.

The announcements were made at Alibaba Cloud's annual Apsara Conference in Hangzhou. The company said it is developing a new AI model that could eventually contain between 5 trillion and 10 trillion parameters. That would make the planned system several times larger than Alibaba's current flagship model.

Parameters are one way of describing the scale of an AI model. They are not a direct measurement of intelligence, but larger models can require enormous amounts of computing power during training and operation. Building a model at this scale therefore requires more than powerful software. Companies also need advanced processors, large data centres, high-speed networking and a reliable supply of electricity.

This is where Alibaba's new chip becomes important. The company introduced the Zhenwu V900, an AI processor developed by its T-Head semiconductor division. Alibaba says the new chip can deliver around three times the performance of its previous generation and is designed to work in large clusters. Mass production is expected to begin in early 2027.

Developing its own processors gives Alibaba another way to reduce its dependence on external chip suppliers. The global AI industry has become heavily dependent on specialised processors, and access to advanced chips has become an important issue for technology companies around the world. Chinese companies in particular have been investing heavily in domestic alternatives as restrictions affect access to some advanced computing hardware.

Alibaba is approaching the AI market from several directions at once. It operates a major cloud business, develops AI models and is now investing in semiconductor technology. The company is also planning to expand its data-centre capacity internationally. Reports say Alibaba is targeting more than 20 gigawatts of data-centre capacity by 2032, showing just how much computing infrastructure the company expects future AI systems to require.

The scale of these plans also shows how expensive the AI race is becoming. Training and operating advanced models requires enormous computing resources. As models become larger and AI applications become more widely used, companies need to keep expanding the infrastructure behind them.

For developers and businesses, this competition could eventually have an impact beyond the hardware industry. More powerful models and cheaper computing could lead to better AI tools for coding, customer support, data analysis, content creation and automation. Cloud providers may also offer more AI services as their infrastructure expands.

At the same time, bigger models do not automatically mean better products. A model's usefulness depends on many factors, including training data, architecture, efficiency, reliability and how effectively it can be integrated into real applications. Companies are therefore competing not only to build larger systems but also to make those systems useful and affordable.

Alibaba's latest announcements are another sign that the AI industry is moving toward a full-stack approach. Instead of depending entirely on third-party hardware and infrastructure, major technology companies are increasingly trying to control multiple parts of the AI pipeline, from chips and data centres to models and cloud services.

The competition is likely to continue as companies in the United States, China and other parts of the world invest heavily in AI infrastructure. For users, developers and businesses, the biggest impact may eventually be seen in the AI services built on top of this technology.

Alibaba's new chip and larger-model plans are still part of a longer development process, but they show where the company believes the industry is heading: toward increasingly large AI systems supported by massive computing infrastructure and a growing amount of specialised hardware.