Global semiconductor sales continue to contract as companies struggle to shake a prolonged slowdown in demand. Sales dropped to $32.9 billion in February from $35.5 billion in January, according to the latest statistics from the Semiconductor Industry Association (SIA). Global sales have decreased more than 10 percent over the last year particularly as the market for memory chips has cooled down.
“Global semiconductor sales fell across the board in February, with all major product categories experiencing drops on a year-to-year and month-to-month basis,” John Neuffer, SIA's president, said in a statement. “Sales were also down across all major regional markets, as the global industry continues to endure a period of slowing sales following record revenues over the last three years.”
Many of the largest suppliers, including Texas Instruments to Micron Technology, have blamed declining demand. Texas Instruments, which acts as a barometer for the semiconductor sector, is struggling for the first time in a long time. Micron plans to reduce its output of DRAM and NAND by around 5 percent this year, moving to cut production as its customers delay orders for memory chips used in servers and smartphones.
Intel's strong growth in data centers has been dashed. Over the last year, the largest companies in cloud computing have burned through billions of dollars building data centers to accommodate all the businesses moving to the cloud. But many of these companies, including Google, Amazon and Microsoft, have started slashing orders for Intel's Xeon processors. They are trying to get through the inventory they already have.
Nvidia's colossal growth has also started to slow down. In February, the company reported fourth-quarter sales totaling $2.21 billion, down 24 percent over the last year and down 31 percent since the third quarter of 2019. The downturn was driven by sluggish sales in its gaming segment and reduced demand for Nvidia's GPUs in data centers, where they are the current gold standard for training artificial intelligence models.