Qualcomm unveils chips to compete with Nvidia, AMD
Make us a preferred source
Make us a preferred source on Google
The news: Qualcomm announced the launch of two AI chips for data centres as it pushes to diversify beyond smartphones while expanding into the ultra-competitive AI infrastructure markets.
The context: The new chips which are called AI200 and AI250, are designed for improved memory capacity and running AI applications, or inference, and will be commercially available in 2026 and 2027, respectively.
Qualcomm’s chips will match Nvidia and AMD, which offer graphics processing units, or GPUs, in full-rack systems linking up to 72 chips to act as one computer. This intensive computing power is needed by AI labs to run advanced models.
Qualcomm emphasised the efficiency of the AI250 chip, underscoring that it will be cheaper for companies in the long term.
A press release published by Qualcomm said: the AI250 will provide “a generational leap in efficiency and performance for AI inference workloads by delivering greater than 10x higher effective memory bandwidth and much lower power consumption. This enables disaggregated AI inferencing for efficient utilization of hardware while meeting customer performance and cost requirements.”
In May Qualcomm flagged its plans to make customer data centre central processing units using Nvidia technology to connect the firm’s AI chips.
To strengthen its AI suite, Qualcomm bought Alphawave in June, which designs semiconductor tech for data centres, for about USD2.4 billion.
Axios reports that Saudi AI company Humain will be the first customer for the new chips, with plans to bring on 200 megawatts of compute starting in 2026.
