When you buy through affiliate links in our content, we may earn a commission at no extra cost to you. Learn 
If you’re buying artificial intelligence chips, chances are they’re going to be made by Nvidia. However, AMD GPUs and CPUs have been catching up.
Not content with being the second largest provider of AI chips, last week AMD made a strategic investment of $655 million in a Finnish AI startup known as Silo AI, one of the largest private AI labs in Europe.
The investment will provide AMD with access to Silo’s 300-person team, which has extensive experience building custom large language models (LLMs), a category of AI models that powers products like ChatGPT.
It is also part of a wider investment strategy which has seen AMD invest over $125 million over the last year, with notable acquisitions including Mipsology and Nod.ai.
Key Takeaways
- This week AMD announced it was investing $655m into Finish AI startup Silo AI.
- AMD hopes to use the investment to bolster its enterprise AI product ecosystem. It is in addition to previous million-dollar investments.
- The performance and cost of AMD chips make them a key competitor in the AI chips market.
- Nvidia remains the world’s largest AI chip maker.
Why The Silo AI Deal Makes Sense for AMD
For AMD, the investment provides an opportunity to produce its own products rather than being confined to the role of a chipmaker. A partnership with the Silo AI team gives AMD access to leading AI engineers with extensive experience building AI models.
“Across every industry, enterprises are looking for fast and effective ways to develop and deploy AI solutions for their unique business needs,” Vamsi Bopping, senior vice president of the Artificial Intelligence Group at AMD in the official release.
“Silo AI’s team of trusted AI experts and proven experience developing leadership AI models and solutions, including state-of-the-art LLMs built on AMD platforms, will further accelerate our AI strategy and advance the build-out and rapid implementation of AI solutions for our global customers,” Bopping said.
It’s worth noting that in 2023, Silo AI generated €21 million in revenue by adding customers including Philips, Rolls-Royce, Nvidia, and Intel. The success of Silo AI’s team is a strong sign that AMD will be able to add some compelling AI-powered products to its product portfolio.
In short, Silo AI provides AMD with more expertise that it can use to enrich its product ecosystem and provide new product offerings to enterprise customers.
AMD’s Place in the AI Chips Market
As demand for AI-powered applications increases, the AI chips market is expected to grow dramatically. Precedence Research estimates that the global AI chips market will grow from a valuation of $16.86 billion in 2022 to $227.48 billion by 2032, growing at a compound annual growth rate of 29.72%.
In this market, AMD has the second largest market share behind Nvidia. While Nvidia claims between 70 to 95% of the market, AMD is comfortable in second place.
On its own terms, AMD is doing pretty well, with AMD’s Chief Executive Officer, Lisa Su reportedly expecting AI chip sales of $4 billion by the end of 2024, but Nvidia’s foothold in the market is a significant roadblock to its long-term growth in the industry.
At the moment, AMD’s CPUs are popular, but many organizations are moving toward Nvidia GPUs, which are specifically optimized to support more intensive AI workloads, capable of training and running AI-powered applications.
AMD Vs Nvidia in the AI Chip Race
When it comes to AI chips, the major choice is between AMD’s Radeon GPUs versus Nvidia’s H100 and H200 GPUs. The real differentiator between the two comes down to performance, i.e. which offers the best model performance at the lowest overall cost of infrastructure.
Oliviar Blanchard, Research Director at The Futurum Group, told Techopedia:
“Right now, Nvidia has the market share upper hand with data center GPUs, with H100 and H200 GPUs essentials being the industry standard for GPUs tasked with handling high-end AI workloads.
“Currently, AMD is a very distant second in terms of Datacenter AI-focused GPU market share, but the company’s new MI300X GPUs are excellent substitutes for Nvidia GPUs, and in some instances even outperform them.”
Blanchard added: “Nvidia’s market penetration advantage notwithstanding, I expect AMD to continue gaining momentum well into 2025 in the data center GPU segment.
“On the data center GPU front, AMD has its EPYC processors, and Nvidia has its Grace “super chips,” but the two products aren’t exactly aligned symmetrically.”
Blanchard argues that in the PC GPU segment, AMD’s Radeon GPUs compete well against Nvidia’s RTX GPUs but suggests that use cases may steer you towards AMD’s overall price and performance or Nvida’s advantages, such as better ray tracing, frame generation, and DLSS.
In this sense, we can see that AMD and Nvidia are both extremely competitive. While Nvidia has the natural market share advantage, AMD does have enough advantages in terms of price and overall performance to be a viable alternative for the enterprise market.
The Bottom Line
AI chips are the fuel that will enable the widespread development and deployment of AI models. AMD has developed a rich product ecosystem, which has all the core ingredients necessary to present a viable challenge to Nvidia’s dominance.
While it’s unlikely that Nvidia will lose its market share anytime soon, AMD GPUs present a much-needed alternative to enterprise investment in AI development.
