Qualcomm Prepares New AI Chip, Meta as First Customer

Qualcomm introduced the Dragonfly C1000 artificial intelligence-based data center CPU, which is scheduled to start production in 2028, with Meta as its first major customer. Photo: Qualcomm

JAKARTA - Qualcomm is starting to get deeper into the AI data center business. The American chip company introduced a new CPU called Dragonfly C1000, with Meta as the first major customer.

Launching Anadolu Agency Thursday, June 25, Dragonfly C1000 is scheduled to begin production in 2028. This chip is designed to handle the workload of agent AI, an AI system that not only responds to commands, but can also perform certain tasks.

Qualcomm said the Dragonfly C1000 is built to deliver high-performance computing with power consumption kept in check. Energy efficiency is now an important factor in data center competition as AI infrastructure requires large amounts of electricity.

The announcement was made in Qualcomm's investor presentation. The company also outlined a broader roadmap for the data center market, including AI chips and products that connect multiple chips.

Qualcomm CEO Cristiano Amon said the company has built a broader portfolio for the next phase in the data center business. According to him, Qualcomm has accumulated assets and strengthened its ability to enter the sector more seriously.

According to an Anadolu Agency report, this move also shows Qualcomm's efforts to reduce its dependence on the smartphone chip business. In the quarter ended March, the mobile phone chip business still accounted for about two-thirds of the company's product revenue.

Qualcomm is also targeting growth in the automotive, robotics, and AI infrastructure sectors. This strategy comes as global smartphone shipments are still below their peak in 2017.

According to Qualcomm, the company's experience designing power-efficient chips for mobile phones and computers can be an important asset to serve hyperscale data center customers. Hyperscale refers to giant data center operators that manage computing on an extremely large scale.

The company also said it had secured two custom silicon chip deals with hyperscale customers.

Amon dismissed concerns that Qualcomm was late to enter the data center market. He said scale, execution, engineering capabilities, operations, and supply chain strength would be the key to competition.

This move puts Qualcomm in the AI data center chip race, a market now being contested by many major technology companies.