According to CCTV Finance citing multiple foreign media reports, AMD (stock price: $134.9; total market value: $218.9 billion) has confirmed plans to lay off approximately 4% of its workforce worldwide, or about 1000 people. However, the company did not disclose the specific departments affected by the layoffs.
Regarding this, AMD stated that layoffs do not mean the company is in financial distress, but rather a refocusing of resources on products with higher profit margins. This move aims to accelerate its artificial intelligence and data center strategy, rather than broader reforms.
An AMD spokesperson stated, "As part of integrating our resources and maximizing growth opportunities, we are taking a series of targeted measures that unfortunately will result in approximately 4% of global layoffs." However, AMD declined to disclose which teams will be downsized.
According to documents from the US Securities and Exchange Commission (SEC), as of the end of last year, AMD had 26000 employees.
AMD's stock price has fallen by over 17% in the past month, with a latest market value of approximately $218.9 billion (about RMB 1584 trillion). AMD's stock price has fallen by 5.5% overall this year, while Nvidia's stock price has risen nearly 200% so far this year.
Although AMD is the world's second-largest GPU manufacturer after Nvidia, Nvidia still firmly controls nearly 90% of the global market in the AI chip field.
To this end, AMD is vigorously developing the research and development of AI chips and investing heavily in it. According to AMD's latest Q3 2024 financial report, the company's R&D investment increased by 9% and total sales costs increased by 11%.
AMD has launched AI chips including the MI300X series, which has also supported the rapid growth of the company's data center business. In the third quarter, Nvidia's data center business revenue increased by 200%. Companies including Meta and Microsoft are trying to use AMD's data center AI chips to reduce their reliance on Nvidia chips.
AMD will launch mass production of the new generation AI chip MI325X series in the fourth quarter. Analysts predict that the overall growth of AMD's data center business will double by 2024, far exceeding the company's 13% increase in total sales.
Last October, AMD predicted that its AI chip sales would reach $5 billion by 2024, accounting for approximately one-fifth of the total projected sales. AMD also predicts that by 2028, the total market size of global AI chips will reach $500 billion.
However, given the limited advanced wafer manufacturing capabilities currently available, increasing the production of AI chips will also face production bottlenecks and require high costs. TSMC remains the world's largest AI chip foundry, producing chips for companies such as NVIDIA and Apple.
According to analysts' predictions, AMD's total sales this year will reach $25.7 billion, while Nvidia's total revenue in 2024 is expected to reach $125.9 billion. Nvidia's scale far exceeds AMD's, which also means that the company's products will receive higher priority in TSMC's production.
On October 29th local time, AMD released its third quarter financial report, which showed that the company's revenue for the period was $6.819 billion, a year-on-year increase of 18%; The net profit was $771 million, a year-on-year increase of 158%. During the period, the company's data center business revenue was 3.549 billion US dollars, a year-on-year increase of 122%, contributing more than half of the company's overall revenue. The data center business has doubled its sales for the second consecutive quarter.
It is reported that AMD's revenue and gross profit margin performance slightly exceeded expectations under non GAAP conditions during the quarter, but its after hours stock price fell on the same day, which is believed to be related to the company's lower than expected Q4 performance guidance.
In the GPU field, the market is increasingly concerned about Nvidia's "monopoly" situation, so AMD's growth is being focused on, and more and more basic large-scale model developers are considering developing their own AI chips. However, in this process, the ecosystem and software and hardware systematization capabilities have become important barriers to the subsequent AI chip market chaos.