For the last thirty years, Nvidia has been the leader in the graphics processing industry, pioneering innovations that have not only shaped the world of gaming, but also professional visualisation, data centre technologies, and even artificial intelligence.
But hot on its heels has been AMD. Standing for ‘Advanced Micro Devices’, AMD was formed back in 1969, making it one of the oldest players in the semiconductor industry, but ever since Nvidia came along in 1993, it has been playing second fiddle.
Despite making significant strides with its Ryzen CPUs and Radeon GPUs – which have steadily closed the performance efficiency gap – AMD has faced numerous challenges in unseating Nvidia as the market leader in graphics, but with its new partnership with Oracle, that all might be about to change.
AMD’s New Partnership
It has to be remembered, of course, that one of Nvidia’s biggest strengths lies in its gaming ecosystem. With years of experience, a vast developer network, and cutting-edge tech like DLSS, Nvidia has built a loyal following among gamers and content creators alike. But as more of those followers get introduced to the iGaming-verse – which has recently surpassed $100 billion in market value, and is still growing – the games they play are shifting away from traditional PC and console titles.
As a result, it’s the other technologies behind Nvidia’s ecosystem that are going to be more important over the next ten years, and this is where AMD can take centre stage. Just recently, for instance, AMD announced a new partnership with Oracle, aimed at expanding OCI – Oracle Cloud Infrastructure – using AMD’s latest high-performance technology.
Reportedly worth billions of dollars, this deal will help Oracle build a cluster of 30,000 MI355X AI accelerators, which are part of AMD’s next-generation Instinct platform. These accelerators are designed specifically for large-scale AI workloads, offering massive memory bandwidth and advanced low-precision computer capabilities, ideal for training and deploying cutting-edge AI models.
More than half of the top ten AI players in the world are already deploying AMD’s data centre chips, but with this partnership, AMD looks set to solidify its position as a formidable force in the AI infrastructure space, potentially even taking over Nvidia as the number one provider.
Will AMD Be Able to Compete With Nvidia?
With artificial intelligence being the fastest-growing segment in tech today, the competition for dominance is more intense than ever. We talked a little before about Nvidia being one of the top providers in the gaming industry, specifically for consoles like Nintendo Switch, but it has to be said, they are just as dominant in the AI space, having held a near-monopoly on AI accelerators over the last decade.
Their GPUs, combined with a mature ecosystem including CUDA and TensorRT, have become the backbone for many AI research and deployment projects, and at the time of writing, they hold around 90% to 95% of the AI accelerator market. So just because AMD has signed this lucrative deal, doesn’t mean they will definitely disrupt the status quo.
But they have a strong chance to. Their accelerators, for instance, offer industry-leading memory bandwidth and energy efficiency, both of which are critical for training large AI models cost-effectively. They’re also built on the advanced CDNA 4 architecture and manufactured on a 3nm process, which means they can support advanced low-precision computing capabilities, which are becoming increasingly important for accelerating AI training and inference without sacrificing any accuracy.
The technology is there to challenge Nvidia, there’s no doubt about that. But at this point, Nvidia’s hold on the market and its own cutting-edge technology might just be enough to keep them at the top for a while longer.