Qualcomm Enters AI Data Center Market With 2 Chips, Takes on Nvidia

Qualcomm will begin selling its AI200 chips next year, with AI 250 chips hitting the market in 2027.

File photo of a Qualcomm building at its San Diego headquarter. Alex Lee/The Epoch Times

Qualcomm Technologies announced on Oct. 27 that it will launch its AI200 and AI250 inference chips and integrated rack-scale systems in the next two years, marking its entry into the AI data center hardware market currently dominated by Nvidia.

The announcement is another strategic pivot for the San Diego-based company, which has predominantly focused on the wireless communications and computing industries and is widely known for its high-performance Snapdragon integrated processors for laptops and tablets. Qualcomm will begin selling its AI200 chips next year, with AI 250 chips hitting the market in 2027, the company said.

Both chips are available in direct-to-chip rack cooling systems, which use significantly less water than other cooling methods deployed in large hyperscale data centers. Qualcomm also noted that it plans to introduce additional AI products to meet the country’s burgeoning data center market.

“We’re redefining what’s possible for rack-scale AI inference,” said Durga Malladi, Qualcomm’s senior vice president and general manager of technology planning, edge solutions and data centers. “These innovative new AI infrastructure solutions empower customers to deploy generative AI at unprecedented [total cost of ownership], while maintaining the flexibility and security modern data centers demand.”

The announcement sent shares of Qualcomm sharply upward in early-day trading, and the stock ended the day rallying more than 11 percent. Qualcomm’s share price has climbed more than 22 percent year to date.

Qualcomm is entering an AI chip and full-rack systems market that’s currently dominated by Nvidia and, to a lesser extent, AMD.

Nvidia’s shares hit new highs on Oct. 27, and its valuation has climbed nearly 39 percent year to date to $4.65 trillion due to insatiable demand for AI computing power. There are more than 5,400 data centers in the United States—more than in any other country—and the industry is expected to grow 12 percent annually through 2030 as artificial intelligence becomes more prevalent.

Globally, companies are expected to invest more than $7 trillion to build out new data centers through 2030, consulting firm McKinsey & Company said. Qualcomm is jockeying for position in the market by focusing on AI chips that run AI models, called inference, rather than chips used in AI training.

Qualcomm said its AI chips support machine learning frameworks, inference engines, generative AI frameworks, and other demanding AI workloads at total costs that are lower than competitors’ due to lower power consumption.

“Our rich software stack and open ecosystem support make it easier than ever for developers and enterprises to integrate, manage, and scale already trained AI models on our optimized AI inference solutions,” Malladi said. “With seamless compatibility for leading AI frameworks and one-click model deployment, Qualcomm AI200 and AI250 are designed for frictionless adoption and rapid innovation.”

Earlier in the year, during President Donald Trump’s visit to Saudi Arabia, Qualcomm announced it had partnered with Saudi-owned AI company Humain to develop and build AI data centers throughout the country using Qualcomm’s AI solutions and data center CPU products.

The collaboration also includes development of semiconductor infrastructure and manufacturing capabilities in the Arab kingdom by the year 2030.