Qualcomm revealed this Monday, the 27th, two new chips aimed at artificial intelligence. This is an attempt by the company to gain space in a sector dominated by NVIDIA.
The AI200 and AI250 are designed for running already trained models, i.e., inference, and not for training neural networks.
The launch represents a strategic shift for the company, historically associated with the manufacture of mobile processors and telecommunications components. Clearly, Qualcomm wants to secure its presence in the data center market and the Artificial Intelligence infrastructure market.
However, it has the challenge of competing with giants like NVIDIA and AMD.
Related News:
Cards on the table
These new processors are based on Hexagon NPU technology, used in the brand’s mobile devices and notebooks. It is evident that they underwent adaptations for large-scale applications. Up to 72 chips can operate together within the same rack.
As a result, they simulate the operation of a supercomputer and configuring a structure similar to that used by NVIDIA GPUs.
The AI200 is scheduled to hit the market in 2026 and will feature 768 GB of LPDDR memory per board. Performance will be optimized for generative and multimodal model inference.
The AI250 is expected to debut in 2027 and will bring a new memory architecture with improved efficiency and lower power consumption.
In theory, he will be able to deliver up to ten times more effective memory bandwidthas well as offering direct liquid cooling, PCIe and Ethernet support for scalability, as well as total power of up to 160 kW per rack.
Possible competition?
NVIDIA dominates the AI ​​chip market, especially with training solutions that are highly sought after by companies developing LLMs. In this sense, Qualcomm’s bet falls on the inference segment.
This is a step that becomes increasingly relevant as generative models become widely used in companies and services.
The new processors are part of Qualcomm’s plan to launch annual generations of AI solutions for data centers. With them, the company reinforces its strategy of offering more efficient and economical alternatives to the options dominated by NVIDIA.
Like this, The company’s focus is on ensuring high inference performance with low power consumption and excellent cost-benefit. For home computers, a 2 Watt difference is irrelevant, but for businesses planning multiple, 2 Watts per rack running 24 hours a day adds up big and fast.
Source: Qualcomm.
Join the Adrenaline offer group
Check out the main offers on hardware, components and other electronics that we found online. Video card, motherboard, RAM memory and everything you need to build your PC. By joining our group, you receive daily promotions and have early access to discount coupons.
Join the group and take advantage of promotions
Source: https://www.adrenaline.com.br/hardware/qualcomm-lanca-chips-de-ia-para-bater-de-frente-com-a-nvidia/
