The third quarter performance report released by global chip giant AMD on Tuesday (October 31) showed that its quarterly revenue exceeded analysts' expectations, but its revenue guidance for the fourth quarter fell short of market expectations.
In the quarter ended September, AMD's adjusted earnings per share were 70 cents, higher than the expected 68 cents; The revenue is 5.8 billion US dollars, a year-on-year increase of 4%, and is expected to be 5.7 billion US dollars.
Meanwhile, AMD expects fourth quarter sales to be approximately $6.1 billion, while analysts expect revenue to be $6.37 billion.
In post market trading, the stock initially fell by about 4%, but after the company made optimistic predictions about its artificial intelligence chip business, the stock subsequently rebounded.
Potential of $2 billion
AMD is one of the few chip manufacturers capable of producing high-end graphics processing units (GPUs), and the market is currently dominated by NVIDIA. GPUs are mainly used for training and deploying generative artificial intelligence models, and the popularity of artificial intelligence has also driven the GPU market this year.
The AMD data center division, which includes server processor CPUs and artificial intelligence chip GPUs, had sales of $1.6 billion in the third quarter, unchanged from the same period last year. AMD expects strong growth in its data center business in the fourth quarter.
AMD CEO Su Zifeng said on the earnings conference call, "We now expect data center GPU revenue to reach approximately $400 million in the fourth quarter and exceed $2 billion by 2024
AMD also mentioned in its performance report that its upcoming artificial intelligence chips MI300A and MI300X will be mass-produced this quarter.
We hope to become an important participant in this market, "said Su Zifeng
During the conference call, Su Zifeng also mentioned its improvements in the enterprise artificial intelligence software suite.
She pointed out, "We have all seen an increase in the workload of generative artificial intelligence, and the fact is that there is still a long way to go before people can truly apply it to enterprise business productivity