Qualcomm is entering the battle for artificial-intelligence computing in the data center.

On Monday, the semiconductor company introduced its AI200 and AI250 rack server lineup, offering new competition in an area dominated by Nvidia and Advanced Micro Devices.

"With Qualcomm AI200 and AI250, we're redefining what's possible for rack-scale AI inference," Qualcomm executive Durga Malladi said in a press.

Qualcomm stock jumped as much as 18% on the news.

The AI200 will be available next year and the AI250 will come out in 2027, according to Qualcomm.

The company said the new servers will incorporate AI accelerator cards that can support up to 768 gigabytes of memory. The memory architecture design will deliver higher memory bandwidth and lower power consumption versus prior generations.

Qualcomm also announced HUMAIN plans to deploy 200 megawatts worth of Qualcomm AI AI200 and AI250 rack solutions in Saudi Arabia and around the world starting next year.

The AI data center market has been moving toward "full stack" rack-scale AI servers which incorporate dozens of chips, networking, and software in one server rack.

Qualcomm is focusing their rack server products for AI inference workloads not AI model training. Inference is the process of generating answers from AI models.

Nvidia is currently shipping rack-scale AI servers including the GB200 NVL72 and GB300 NVL72. AMD plans to ship their first rack-scale AI server, the MI400, next year.