Is Nvidia’s control stifling creativity in the GPU sector?
Is Nvidia’s control stifling creativity in the GPU sector?
Nvidia continues to lead strongly in consumer and AI GPU segments, raising questions about whether this dominance hinders progress or accelerates it. With AMD advancing through RDNA3 and Intel’s Arc showing gradual gains, Nvidia still sets pricing trends and shapes software environments such as DLSS and CUDA. As a community of enthusiasts and builders, we’re curious: are we experiencing a golden era for GPUs or a situation where competition is limited? What do you think about AMD and Intel’s potential to disrupt the landscape? Does Nvidia’s performance justify its market position based on actual gaming and productivity benefits? Let’s explore the future direction of the market and assess whether genuine competition remains.
Nvidia remains the clear leader without doubt. Yet, AMD presents an interesting case with their latest 9070/9070xt models. Although they’re difficult to locate at the MSRP, this trend is consistent across the industry. I believe AMD actually surpassed Nvidia in sales of these units compared to their entire 50 series.
In contrast, AMD is currently a minor player. The 9070xt offers performance similar to the 5070ti, but at $150 less than Nvidia’s MSRP. The 9070 appears to hold an advantage over the 12GB 5070, and the upcoming 9060xt with 16GB is expected to outperform the 5060ti.
Additionally, there are rumors about Intel developing new GPUs. While Nvidia isn’t necessarily retreating, there are whispers of a potential 5080 super and other releases. If these new products gain traction, why would they choose to refresh after just over a year?
In summary, Nvidia maintains dominance while AMD and Intel strive to close the gap. Nvidia’s focus on AI creates a promising avenue for growth.
Nvidia is making significant efforts this year with the 5000 series to improve performance. There are still issues with ROPs, load line calibration on the 12v plug to prevent overheating, driver problems, and no reviews for the 5060/ti model until the official launch. There are also concerns about how reviewers are being treated and the potential for a reviewer blacklist. Specific models include 5060 8GB, which is dropping to a 5050 variant for better performance than the CPU graphics; 5060ti 8GB, which hasn't been released; 5070 12GB with a 16GB variant; 5060ti 16GB, available in 5060 or 8GB versions; 5070ti 16GB leading to the 5080; 5080 with 16GB, followed by 5090 at 32GB; and a Titan model for 5090. The 2026 release is expected to bring Ti versions or upgraded models. If AMD ever successfully launched their 2017 Ryzen GPUs, it would be impressive. Between 2010 and 2017, AMD focused on the K10 and Bulldozer lines until Ryzen's debut, showing confidence in their 14nm+ core CPUs. I’m hoping the CPU industry will begin sharing technology through chiplets and 3D v-cache with GPUs, encouraging AMD to improve. Honestly, the market needs more coordination—CPUs typically release every few years while GPUs follow an even schedule. We still have workstations running Intel i5 2500 CPUs, and upgrades are mainly driven by the inability to run Windows 11. Many people still use Ryzen 5000 CPUs, and since then there have been multiple CPU generations and two more for CPUs.
Depends on the market perspective you consider.
Gaming cards represent a minor portion of their overall income.
Data Center 115 Billion
Gaming 11 B
Work station 2 B
For gaming, there is little competition in the premium segment, though this accounts for only a small share of sales.
Mid-range products dominate the sales figures.
Their current weakness lies in driver problems. I don’t recall them ever facing such issues before, nor do I see a pattern of neglecting driver readiness for gaming cards, given their limited revenue impact.
From what I see, AMD seems unmotivated to capture more market share. They might boost sales by adopting 256-bit cards, but this would likely appeal mainly to enthusiasts who already understand what matters, a niche audience.
Even though the work AMD has done with the 9070 series is remarkable, I believe they intend to move on to UDNA next. This might mark a new era focused on rebuilding and gradually improving the software aspects and performance scaling.
The main issue for others is that many believe Nvidia is superior. I was really excited when Intel entered the GPU market. Great! They’ve spent years developing CPUs with graphics, so they can easily adapt that expertise to create dedicated GPUs. Of course, it’s not that simple. Personally, if I owned a Nvidia, the drivers would be my biggest concern—they usually work out eventually. Not here, but sometimes people defend Nvidia in a very extreme way. The biggest companies in the world don’t care about you. Not that billion-dollar firms do, but better than the worst of them.
It’s definitely affecting pricing. The tricky part is that AMD has the technical skills to compete, but they haven’t managed to maintain a strong position over the past ten years.
But we understood Zen would arrive, bulldozer wasn’t a temporary pause—it was simply postponed by two years. If it had arrived on time in 08 instead of 2010, the situation would have been completely different. Bulldozer suffered delays, which made it less competitive and limited its growth potential due to a shorter development path. I truly believe Lisa Sus’ leadership and her engineering expertise recognized the business needs, which ultimately led to the right direction. amD had strong AI GPUs with real instinct, but they deliberately omitted a Halo product and stopped pursuing Nvidia to capture market share as they did in their early CPU days. It really only mattered having a superior product compared to Intel. … They also abandoned the NAVI 43 project, leaving us wondering where the actual Zen product is being developed. … It’s impressive they’re offering a mid-tier solution almost as good as MediaTek, but the absence of vision and Halo strategy weakens their position in this market. Even with competitive AI GPUs, they’re still trailing Nvidia. What does this AMD GPU approach really mean?