AMD unveiled a new artificial intelligence chip on Thursday that directly targets Nvidia’s data center graphics processors, known as GPUs.
The Instinct MI325X, as the chip is called, will go into production before the end of 2024, AMD said Thursday during an event announcing the new product. If AMD’s AI chips are seen by developers and cloud giants as a close substitute by Nvidia products, could put price pressure on Nvidia, which has enjoyed roughly 75% gross margins while its GPUs have been in high demand over the past year.
Advanced genetic AI, such as OpenAI’s ChatGPT, requires massive data centers full of GPUs in order to do the necessary processing, which has created demand for more companies to provide AI chips.
In recent years, Nvidia has dominated the majority of the data center GPU market, but AMD has historically been a close second. Now, AMD is aiming to take share from its Silicon Valley rival, or at least capture a large slice of the market, which it says will be worth $500 billion by 2028.
“AI demand has actually continued to take off and is actually exceeding expectations. It’s clear that the pace of investment continues to increase across the board,” AMD CEO Lisa Su said at the event.
AMD didn’t reveal new major cloud or web customers for its Instinct GPUs at the event, but the company previously revealed that both After and Microsoft buy its AI GPUs and that OpenAI uses them for some applications. The company also did not disclose pricing for the Instinct MI325X, which is usually sold as part of a complete server.
With the launch of the MI325X, AMD is accelerating its product schedule to release new chips on an annual schedule to better compete with Nvidia and take advantage of the AI chip boom. The new AI chip is the successor to the MI300X, which started shipping late last year. AMD’s 2025 chip will be called the MI350, and the 2026 chip will be called the MI400, the company said.
The launch of the MI325X will pit it against Nvidia’s upcoming Blackwell chips, which Nvidia said will start shipping in significant quantities early next year.
A successful launch of AMD’s newest data center GPU could attract interest from investors looking for additional companies to benefit from the AI boom. AMD is only up 20% so far in 2024, while Nvidia stock is up over 175%. Most industry estimates say Nvidia has more than 90% of the market for data center AI chips.
AMD stock fell 4% on Thursday. Shares of Nvidia rose about 1%.
AMD’s biggest obstacle to gaining market share is that its rival’s chips use its own programming language, CUDA, which has become a standard among artificial intelligence developers. This essentially locks developers into Nvidia’s ecosystem.
In response, AMD said this week that it is improving its competing software, called ROCm, so that AI developers can easily switch more of their AI models to AMD’s chips, which it calls accelerators.
AMD has framed its AI accelerators as more competitive for use cases where AI models create content or make predictions than when an AI model processes terabytes of data for improvement. That’s partly due to the advanced memory AMD uses in its chip, he said, which allows it to serve Meta’s Llama AI model faster than some Nvidia chips.
“What you see is that the MI325 platform offers up to 40% more inference performance than the H200 in Llama 3.1,” Su said, referring to of Meta AI model in big language.
Also taking on Intel
While artificial intelligence accelerators and GPUs have become the most closely watched part of the semiconductor industry, AMD’s main business has been central processing units, or CPUs, which are at the heart of nearly every server in the world.
AMD’s data center sales in the June quarter more than doubled last year to $2.8 billion, with artificial intelligence chips accounting for only about $1 billion, the company said in July.
AMD takes about 34% of the total dollars spent on data center CPUs, the company said. This is still less than Intelwhich remains the market boss with the Xeon series of chips. AMD aims to change that with a new line of CPUs, called EPYC 5th Gen, that it also announced on Thursday.
These chips come in a number of different configurations ranging from a low-cost, low-power 8-core chip that costs $527 to 192-core, 500-watt processors intended for supercomputers that cost $14,813 per chip.
The new CPUs are particularly good at feeding data into AI workloads, AMD said. Almost all GPUs require a CPU in the same system to boot the computer.
“Today’s artificial intelligence is really about CPU capability, and you see that in data analytics and a lot of these types of applications,” Sue said.