"Storage shortage" intensifies! Intel's Chen Liwu: The shortage situation may continue until 2028

Intel CEO Patrick Gelsinger recently predicted that the shortage of storage chips will last at least another two years.

On Tuesday, Gelsinger spoke about the storage chip shortage at the AI summit hosted by Cisco Systems in San Francisco. He stated, “We won’t see relief until 2028.”

The booming artificial intelligence industry has driven rapid expansion of data centers, leading to a surge in demand for storage chips, which in turn has caused supply shortages and price increases. As the world’s largest manufacturer of personal computer processors, Intel also heavily relies on memory chips for data storage and management.

Gelsinger forecasted that NVIDIA’s new generation AI accelerator—the Vera Rubin platform—will further boost demand for storage chips, which he said “will consume a massive amount of storage chips.”

This assessment indicates that, in the storage chip market dominated by Samsung Electronics, SK Hynix, and Micron Technology, the supplier advantage will remain stable.

As global semiconductor industry executives grow increasingly concerned about storage chip supply shortages, analysts expect the seller’s market led by Samsung Electronics and SK Hynix to continue.

Just as Intel issued this warning, the U.S. stock storage sector remained hot on Tuesday. Western Digital’s approval of an additional $4 billion buyback program pushed shares up 7.4%, and SanDisk rose 4.55%, both hitting new highs.

Leading companies in the global EDA and semiconductor IP fields, Synopsys, CEO Sassine Ghazi, recently stated in an interview that the chip “shortage” phenomenon will persist through 2026 and 2027.

According to the latest storage industry survey by TrendForce, in the first quarter of 2026, the continued growth in AI and data center demand will exacerbate the global storage supply-demand imbalance. Original equipment manufacturers’ bargaining power will continue to increase. Based on this, TrendForce has revised upward the quarterly growth rates of prices for DRAM and NAND Flash products in the first quarter. The overall contract price for conventional DRAM is expected to increase from the initial forecast of 55-60% to 90-95%, and NAND Flash contract prices from 33-38% to 55-60%, with further upward potential.

Entering the GPU Market

Gelsinger also announced that Intel will enter the graphics processing unit (GPU) market, which is currently dominated by NVIDIA. This move is seen as part of Intel’s strategic plan: to strengthen its competitiveness in AI semiconductors through self-designed GPUs, and to revive its former glory by leveraging its foundry business.

“We recently recruited a highly capable GPU design leader,” Gelsinger revealed in an interview, though he did not disclose the person’s name. It is understood that this leader is Eric Demers, a GPU expert recruited from Qualcomm last month. Demers announced on LinkedIn that he had joined Intel as Senior Vice President.

Gelsinger said that the GPU development project will be led by Kevork Kechichian, Vice President of Intel’s Data Center Group, who joined Intel last year from Arm. “GPUs are closely related to data center business,” Gelsinger said, “and we will work with customers to clarify their needs.”

GPUs are core semiconductors used in AI development and data centers, with NVIDIA products occupying over 80% of the AI accelerator market share. Major AI chip manufacturers, including NVIDIA, mostly outsource production to foundry TSMC.

Gelsinger also stated that, leveraging new manufacturing technologies, Intel’s foundry business is progressing steadily. “Several customers are actively collaborating with Intel’s manufacturing division,” he added. The main focus of customer interest is Intel’s 14A process technology—a cutting-edge process with a 1.4-nanometer manufacturing precision.

(Article source: Caixin)