Qualcomm has officially announced the introduction of its rack-level artificial intelligence accelerators: the Qualcomm AI200 and AI250 chips, along with corresponding rack solutions, directly challenging NVIDIA's dominance in the field. Following this announcement, Qualcomm's stock price experienced a notable surge.

Both the AI200 and AI250 leverage Qualcomm's proprietary NPU technology, with a strong emphasis on high memory capacity and cost efficiency. The AI200 is projected to hit the market in 2026, primarily targeting AI inference tasks. Meanwhile, the AI250 is slated for a 2027 release and will introduce a groundbreaking new memory architecture. This move signifies Qualcomm's accelerated strategic shift, marking its official entry into the data center AI chip market.

Additionally, Qualcomm revealed plans to launch a new computational chip on an annual basis, although specific pricing and foundry details remain undisclosed. Presently, Qualcomm's revenue streams are predominantly derived from its semiconductor and technology licensing operations. The introduction of these AI chips is strategically aimed at fostering a fresh avenue for revenue growth. Considering NVIDIA's current stronghold in the AI chip market, Qualcomm is poised to encounter even more intense competition in the foreseeable future.