AMD launched its latest MI325X AI chip designed to compete with Nvidia Blackwell in the fierce competition of the 2024 AI chip market. What are its features and advantages? AMD launches its MI325X AI chip, designed to rival Nvidia’s Blackwell, promising superior performance in AI applications and deep learning workloads. AMD launches a new artificial-intelligence chip on Thursday that is taking direct aim at Nvidia’s data center graphics processors, known as GPUss The Instinct MI325X, as the chip is called, will start production before the endMI325X of 2024, AMD launches said Thursday during an event announcing the new product.
In October 2024, AMD officially launched its latest MI325X AI chip, a major innovation that is expected to be a serious competitor to Nvidia, especially with their Blackwell series. Amidst the fierce competition in the AI chip industry, AMD is trying to present a new solution that can improve the performance of artificial intelligence (AI) and machine learning (ML) with cutting-edge technology. In this article, we will explore the features, performance, and impact of the MI325X and how this chip competes with Nvidia Blackwell, which has dominated the market for the past few years.
Features and Advantages of AMD MI325X
The AMD MI325X chip is specifically designed to accelerate AI and machine learning computing. With the support of a more efficient architecture, the AMD MI325X relies on high speed and greater computing power, while maintaining energy efficiency. With these specifications, this chip is designed for use in large-scale systems that require large amounts of data processing such as data centers, supercomputers, and AI-based cloud services.
One of the main advantages of the MI325X is the use of more advanced multi-core technology compared to the previous generation. This allows faster data processing and supports more complex applications. In addition, this chip is also equipped with optimal tensor core support for AI processing, similar to the Nvidia architecture which also uses tensor core technology.
AMD MI325X AI chip vs Nvidia Blackwell: Who is Superior?
When comparing AMD MI325X with Nvidia Blackwell, the main question is: who has the best performance to handle large-scale AI applications? Nvidia has been known as the market leader in AI GPUs with the Blackwell architecture which is recognized for its power and energy efficiency. However, AMD has shown that the MI325X has several aspects that make it very competitive.
The graphics and processing performance of the MI325X are said to be close to, if not equal to, Nvidia Blackwell. In some early benchmarks released, AMD showed that the MI325X has processing speeds comparable to Nvidia in certain tasks such as neural network training and AI inference. Additionally, AMD highlighted the MI325X’s power efficiency, which can save up to 20% energy compared to its competitors, making it an attractive option for data centers that need higher efficiency.
If AMD’s AI chips are seen by developers and cloud giants as a close substitute for Nvidia’s products (as rival), it could put pricing pressure on Nvidia, which has enjoyed roughly 75% gross margins while its GPUs have been in high demand over the past year.
Advanced generative AI such as Open AI’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 the past few years, Nvidia has dominated the majority of the data center GPU market, but AMD is historically in second place. Now, AMD is aiming to take share from its Silicon Valley rival or at least to capture a big chunk of the market, which it says will be worth $500 billion by 2028.
“AI demand has actually continued to take off and actually exceed expectations. It’s clear that the rate of investment is continuing to grow everywhere,” AMD CEO Lisa Su said at the event.
AMD didn’t reveal new major cloud or internet customers for its Instinct GPUs at the event, but the company has previously disclosed that both Meta 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 typically 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 boom for AI chips. The new AI chip is the successor to the MI300X, which started shipping late last year. AMD’s 2025 chip will be called MI350, and its 2026 chip will be called MI400, the company said.
The MI325X’s rollout will pit it against Nvidia’s upcoming Blackwell chips, which Nvidia has said will start shipping in significant quantities early next year.
A successful launch for AMD’s newest data center GPU could draw interest from investors that are looking for additional companies that are in line to benefit from the AI boom. AMD is only up 20% so far in 2024 while Nvidia’s stock is up over 175%. Most industry estimates say Nvidia has over 90% of the market for data center AI chips.
AMD launches stock fell 3% during trading on Thursday.
AMD’s biggest obstacle in taking market share is that its rival’s chips use their own programming language, CUDA, which has become standard among AI developers. That essentially locks developers into Nvidia’s ecosystem.
In response, AMD this week said that it has been improving its competing software, called ROCm, so that AI developers can more easily switch more of their AI models over to AMD’s chips, which it calls accelerators.
AMD has framed its AI accelerators as more competitive for use cases where AI models are creating content or making predictions rather than when an AI model is processing terabytes of data to improve. That’s partially due to the advanced memory AMD is using on its chip, it said, which allows it to server Meta’s Llama AI model faster than some Nvidia chips.
“What you see is that MI325 platform delivers up to 40% more inference performance than the H200 on Llama 3.1,” said Su, referring to Meta’s large-language AI model.
Taking on Intel, too
While AI accelerators and GPUs have become the most intensely watched part of the semiconductor industry, AMD’s core business has been central processors, or CPUs, that lay at the heart of nearly every server in the world.
AMD’s data center sales during the June quarter more than doubled in the past year to $2.8 billion, with AI chips accounting for only about $1 billion, the company said in July.
AMD takes about 34% of total dollars spent on data center CPUs, the company said. That’s still less than Intel, which remains the boss of the market with its Xeon line of chips. AMD is aiming to change that with a new line of CPUs, called EPYC 5th Gen, that it also announced on Thursday.
Those chips come in a number of different configurations ranging from a low-cost and 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 for feeding data into AI workloads, AMD said. Nearly all GPUs require a CPU on the same system in order to boot up the computer.
“Today’s AI is really about CPU capability, and you see that in data analytics and a lot of those types of applications,” Su said.
Technology Behind MI325X AI chip
To provide greater competitiveness, AMD uses the latest CDNA3 architecture on the MI325X, which allows for increased computing performance while reducing latency. This technology is also combined with better parallel processing capabilities, allowing the MI325X to process more data in less time.
In addition, AMD’s Infinity Fabric allows for better connectivity between the CPU and GPU, making the overall system performance smoother. This technology also increases memory bandwidth, which is important for AI applications that require fast access to large amounts of data. Nvidia may have an advantage in its more mature CUDA architecture, but AMD has managed to leverage the full integration between the CPU and GPU to accelerate processing.
AI Chip Market and Impact of AMD-Nvidia Competition
The launch of AMD MI325X is not only eye-catching from a technological perspective, but also from a market perspective. In recent years, Nvidia has dominated the AI chips market, especially with its popular A100 and H100 series. However, AMD with MI325X seems to want to change the landscape. The improved performance, energy efficiency, and better integration with AMD’s overall system make the MI325X an attractive alternative for companies looking for a more affordable yet high-performance AI solution.
AMD often offers competitive pricing compared to Nvidia, so companies looking for cost-effectiveness may prefer AMD’s solution. On the other hand, Nvidia has a well-established ecosystem, including AI software and frameworks optimized for its GPUs, such as NVIDIA CUDA and TensorRT. Therefore, the battle between the MI325X and Blackwell is not just about hardware, but also the supporting ecosystem.
AMD’s Future Potential
If AMD can maintain the innovation trend seen with the MI325X, then going forward, the company has a great opportunity to displace Nvidia’s dominance in the AI chip market. AMD has shown the ability to compete with Nvidia, not only in terms of GPU performance for gaming, but now also in more demanding categories, such as AI and machine learning.
One thing to watch is how AMD positions the MI325X in the market, whether they will focus on the enterprise segment or try to expand adoption among general users interested in AI development. If AMD can successfully build a stronger software ecosystem to support the MI325X, they could become a major player in the near future.
AMD’s Challenges Against Nvidia
However, it is undeniable that Nvidia has a big advantage in terms of their AI ecosystem. Nvidia has invested heavily in the software and AI frameworks that support their GPUs. With NVIDIA AI Enterprise, they have built a solid AI platform that makes it easy to develop and deploy AI applications at scale.
AMD needs to be more aggressive in developing a software ecosystem that supports the MI325X. Without strong software support, hardware performance alone may not be enough to beat Nvidia’s dominance. However, if AMD can build a more inclusive and easy-to-use platform, they can attract more AI developers who previously focused on Nvidia.
Conclusion
The AMD MI325X is a significant step in AMD’s strategy to challenge Nvidia’s dominance in the AI chip market. With superior features such as higher power efficiency, competitive performance, and a potentially more affordable price, the MI325X offers an attractive alternative for companies looking for a high-performance AI chip. However, the biggest challenge facing AMD is the software ecosystem and developer support that Nvidia has long dominated.
Going forward, the battle between AMD MI325X and Nvidia Blackwell will be one of the main stories in the AI technology industry. Can AMD shake Nvidia’s dominance? Only time will tell, but with the MI325X, AMD has shown that they are no longer a player to be ignored.