News
AMD AI Chip Challenge Fails to Boost Stock Amid 5% Plunge
Despite launching a new product that claims to outperform NVIDIA's similar artificial intelligence (AI) chips in terms of reasoning capabilities, AMD's stock price still ended with a significant drop. It appears that investors are not impressed by verbal presentations and are truly looking forward to signs of returns on AI investments.
On Thursday, October 10th, Eastern Time, during AMD's data center event where they introduced the MI325X AI accelerator and several network chips, AMD's stock price only briefly turned positive at the end of the early morning session and the beginning of the midday session. After that, it maintained a downward trend throughout the midday session, reaching a daily low of $162 at the end of the day, with the daily decline expanding to about 5.3%, and ultimately closing down by 4%. This marked two consecutive days of decline to a weekly low, creating the largest intraday and closing decline since September 3rd.
One would think that the introduction of a powerful AI chip should be a positive development, especially since AMD's new product is targeted at NVIDIA's highly anticipated Blackwell architecture chips. So, why is AMD's stock price accelerating downward?
Some media have summarized that the stock price decline indicates that investors are still waiting for AMD's AI business to generate returns. Their commentary suggests that although AMD has released new chips, the company's financial outlook for the near term has not changed significantly. Like NVIDIA, AMD has also promised to release new AI accelerators annually, accelerating the pace of innovation. However, AMD has a long way to go to catch up with NVIDIA, and Wall Street has been waiting for signs of progress in AMD's pursuit of NVIDIA. This progress may only be visible after AMD's next financial report is released, which is expected around the end of this month when AMD publishes its financial results for the third quarter of this year.
Advertisement
The financial report released at the end of July this year showed that AMD's data center business net revenue was $2.8 billion, a year-on-year increase of 115%. This was mainly due to the strong demand for the AI accelerator MI300. As the flagship product competing with NVIDIA's H100, MI300's quarterly sales exceeded $1 billion, far exceeding market expectations. At that time, AMD CEO Lisa Su stated that AMD's AI chip sales were "higher than expected," and the company planned to launch the MI325X in the fourth quarter of this year, and next year to launch the MI350, which should be "very competitive" compared to NVIDIA's Blackwell, with the MI400 to be launched in 2026.
After AMD's financial report was released, some analysts believed that NVIDIA still has a leading advantage over AMD. AMD's data center business scale is still small compared to NVIDIA's - with only $2.8 billion in sales in one quarter. In contrast, NVIDIA's data center revenue for the first fiscal quarter of the 2025 fiscal year, which ended in April, was $22.6 billion, setting a new record for the company's single-quarter business, with a sequential growth of 23% and a year-on-year surge of 427%.
At this Thursday's data center event, AMD introduced that the company's latest AI data center GPU, the MI325X, is expected to ship in the fourth quarter of this year and will be launched in the first quarter of next year through products搭载ing this GPU from AMD partners such as Dell, Eviden, Gigabyte, Huiyu, Lenovo, and Supermicro.
Some media have reported that Thursday's event reflects AMD's increasing optimism about the long-term market size that AI chips may reach. Lisa Su said that over the past year, AI demand has exceeded AMD's expectations, and she now estimates that by 2028, the market size for AI data center GPUs will reach $500 billion, with an annual growth rate of over 60%. In December last year, she estimated that the market would break through $400 billion by 2027.
As for whether AMD's MI325X will defeat NVIDIA's products, the media cannot assert this, implying that it is still an unknown. They point out that the most comparable to MI325X is not NVIDIA's H200, but NVIDIA's most powerful Blackwell architecture GPU. The Blackwell GPU has begun to be introduced to customers this quarter, and its performance may be superior to MI325X, which is based on the same CDNA 3 architecture as AMD's older MI300X GPU.
Wall Street Journal mentioned earlier that Lynx Equity Strategies commented that Thursday's significant drop in AMD's stock price may stem from the fact that MI325X cannot significantly change AMD's lagging status and cannot effectively challenge NVIDIA's dominant position. "Unless AMD can demonstrate a clear driving force for market share growth, the stock may retest its annual low."Lynx Equity Strategies also noted that AMD's "management may tout benchmark test results of its chips as superior to Nvidia's Hopper series, but in actual workload operation, we believe the MI300X still lags behind Nvidia's H100. In addition to the exceptional HBM memory density, the software stack, backplane network, and power consumption are equally important."
However, KeyBanc analyst John Vinh believes that even if AMD is far behind Nvidia, AMD's development trajectory is clearly optimistic, with an expected shipment of 500,000 Instinct MI300X AI accelerators this year, gaining an increasingly larger share in the AI market.
Additionally, some media have reported that the AI hardware sector is attracting substantial budgets from companies, with AMD hoping to get a slice of the pie. Nvidia is currently the leader in the entire market, and some analysts believe that, given the huge opportunities in the AI hardware sector, there is room for multiple companies to win.
Leave A Comments