AMD Verifies Withdrawal from Enthusiast GPU Segment in Order to Concentrate on Increasing Market Share

AMD stated in a Tom’s Hardware interview that the enthusiast graphics market will not be the focus of its upcoming RDNA 4 graphics architecture-based gaming GPUs. In an interview with Paul Alcorn, Jack Huynh, head of AMD’s Computing and Graphics Business Group, stated that the company’s goal for its upcoming generation is to capture market share in the PC gaming graphics market by outperforming NVIDIA in price-performance battles in key mainstream and performance segments, much like it did with the Radeon RX 5000 series, which was built on the original RDNA graphics architecture. AMD will also avoid entering the low-margin enthusiast segment, which is driven by low volume sales and low die-sizes. AMD desperately needs to improve its current 12% market share in gaming discrete GPUs, especially considering that its graphics intellectual property is up to date.

Huynh responded, “I am looking at scale, and AMD is in a different place right now,” to a pointed question about whether AMD will continue to cater to the enthusiast GPU market given that funding for cutting-edge wafers is best spent on data-center GPUs. I believe we discuss this topic a lot at AMD. I would like to know if you believe that the PlayStation 5 is harming us. It costs $499. So tell me, is playing King of the Hill enjoyable? Once more, I’m seeking scale. Because I bring developers along when we reach size. Building scale is therefore my top priority at the moment in order to get us closer to 40–50 percent of the market. Should I aim for 10% or 80% of the Total Addressable Market? I don’t want AMD to be a firm that is exclusive to those with the means to purchase Ferraris and Porsches, so I’m an 80% sort of man. Millions of people will use the gaming platforms we hope to develop. Indeed, we will have incredibly fantastic products. However, we attempted the King of the Hill tactic, and it didn’t really take off. After ATI tried this King of the Hill tactic, its market share essentially remained the same. My goal is to create the greatest items at the appropriate system cost. Thus, consider pricing points; we’ll have a leader.”

Huynh responded, “One day, we may,” to Alcorn’s pressing question, “Price-wise, you have leadership, but you won’t go after the flagship market?” However, building AMD’s size is my top goal at the moment. I can’t acquire the developers right now without scale. When I tell developers that my goal is only to capture 10% of the market, they simply respond with, “Jack, good luck, but we have to go with Nvidia.” Therefore, I have to present them with a plan that says, “Hey, this strategy will help us get to 40% market share.” ‘I’m with you now, Jack,’ they then say. I’ll now optimize for AMD. We can pursue the top after we have it.

The exchange appears to attest to AMD’s plan to leave the enthusiast market, citing low volumes for the type of technical work and high wafer costs required to produce GPUs for this market. Radeon RX 6800 and RX 6900 series were quite successful for the firm, mostly because the RDNA 2 generation took advantage of the need for high-end GPUs due to the GPU-accelerated cryptomining frenzy. By the time AMD released its next-generation Radeon RX 7900 series powered by RDNA 3, this demand had vanished, and the company’s prospects were negatively impacted by the lack of performance leadership when compared to the GeForce RTX 4090 and RTX 4080 with ray tracing enabled. Reports that AMD is concentrating on the performance sector (and below) are consistent with the speculation that, in order to lessen the performance impact of enabling ray tracing, AMD is actively working to improve its ray tracing performance with RDNA 4. Raster performance and efficiency may be the company’s strategies for expanding its market share.

AMD’s major premise here is that it has a product issue rather than a distribution issue, and that it will increase its market share with a product that balances the performance/Watt and performance/price equations.