“We are taking a number of steps to reduce our global workforce by approximately 4% to focus our resources on our biggest growth opportunities,” an AMD spokesperson said, according to CRN. “We will respect and support affected employees to help them navigate this transition.” However, the departments that will be mainly affected by the cuts have not yet been clearly identified.
We requested additional explanation from AMD, but there was no response.
These layoffs surprised the industry as they were announced amidst strong quarterly performance, including significant increases in AMD’s sales and net profit. Employees also could not hide their surprise as information related to layoffs was first revealed on the community chat platform Blind, and the company later confirmed it.
However, a closer look at the performance reveals a mix of strengths and challenges. AMD’s third quarter sales increased 18% to $6.8 billion, but gaming chip sales plummeted 69% compared to the previous year, and embedded chip sales also fell 25%.
AMD CEO Lisa Su emphasized in a recent earnings announcement that the data center and AI business are key to the company’s future, and that 98% growth is expected in this sector by 2024. Sue cited orders from customers such as Microsoft and Meta as a factor in the recent increase in sales. In particular, Meta introduced AMD’s MI300X GPU to its internal workload.
Unlike AMD’s relatively targeted, small-scale reductions, Intel recently implemented much larger-scale job cuts, including eliminating approximately 15,000 jobs as part of its restructuring.
Data centers, AI, AMD’s growth drivers
AMD is growing rapidly by optimizing its Instinct GPUs for AI workloads and meeting data center reliability standards. As a result, the sales forecast for the Instinct product line in 2024 was raised by $500 million.
Major customers such as Microsoft and Meta are also expanding their use of MI300X GPU. Microsoft is applying this to its Co-Pilot service, and Meta is applying it to the Llama model. Additionally, public cloud providers such as Microsoft and Oracle Cloud and several AI startups are also adopting MI300X instances.
This focus on AI resulted in AMD’s R&D costs increasing by approximately 9% in the third quarter. This is to support expanded production of the MI325X AI chip, which is scheduled to be released later this year.
In addition, AMD recently introduced its first open source large-scale language model under the OLMo brand, establishing a strong foothold in competition with NVIDIA, Intel, and Qualcomm in the AI market.
“AMD has the potential to build a complete AI solution across hardware, LLM and the broader ecosystem,” said Sushil Menon, director, Everest Group. “This will be a point of differentiation for AMD among the major silicon manufacturers. “You can,” he said.
AMD is considered Nvidia’s closest competitor, supporting advanced data centers that handle the massive data demands of generative AI technology.
editor@itworld.co.kr
Source: www.itworld.co.kr
