The wait is over, and AMD has finally unveiled their Radeon 97 series GPUs. After months of speculation and leaked retail packaging sightings, these new graphics cards appear poised to shake up the market in meaningful ways. What’s particularly striking about this launch isn’t just the technical specifications – it’s the possibility that we might actually be able to buy one.
The new RDNA 4 architecture marks a significant departure from AMD’s recent approach, returning to a monolithic die design. At $599, the flagship model offers 16GB of VRAM, enhanced media capabilities, third-generation ray tracing accelerators, and second-generation AI accelerators. This combination of features at this price point could finally give NVIDIA some real competition.
Performance Improvements That Matter
The numbers AMD is presenting are genuinely impressive. Each compute unit is reportedly 40% faster than its RDNA 3 predecessor, thanks to improved scalar units and dynamic register allocation. This isn’t just an incremental update – it’s a complete architectural overhaul that could reshape the mid-range GPU market.
Ray tracing capabilities have seen substantial improvements, with AMD adding:
- A second ray intersection engine for doubled throughput
- Dedicated hardware for ray transformation
- New oriented bounding boxes accelerator
While these improvements might not fully match NVIDIA’s ray tracing prowess, they represent a significant step forward for AMD. The gap between the two manufacturers in this crucial technology appears to be narrowing.
AI and Media Features That Add Real Value
The AI capabilities of these new cards show promising potential. AMD claims a 2x performance improvement in FP16 operations and 4x in INT8 compared to RDNA 3. What’s particularly interesting is how AMD is implementing AI features that could actually benefit everyday users, rather than just chasing benchmark numbers.
The new features include:
- Image Inspector: Monitors game frames for corruption during gameplay
- AMD Chat: A local large language model for GPU configuration and tuning
- Enhanced media engine with 8K/60 FPS encode and decode capabilities
- Unrestricted encoding sessions – unlike NVIDIA’s artificial limitations
The Price-Performance Equation
The pricing strategy for the 9770 and 9770 XT reveals AMD’s only apparent misstep in this launch. With just a $50 difference between models showing significant performance variations, the value proposition becomes muddied. The 9770 XT’s 84-watt higher power draw and substantially better specifications make the base 9770 seem like an afterthought at its current price point.
AMD should consider repricing the base 9770 to $499 to create a clearer market segmentation and avoid the mistakes made with the 7900 XT launch.
Availability: The Make-or-Break Factor
For the first time in as long as I can remember, AMD might not have screwed up a GPU launch.
The most promising aspect of this launch isn’t the technical specifications or features – it’s the apparent readiness for mass distribution. These cards have been shipping to retailers for months, suggesting AMD might actually deliver on their promise of wide availability starting March 6, 2025.
The timing couldn’t be better. With NVIDIA’s latest cards being scarce and Intel’s budget offerings hard to find at MSRP, AMD has a unique opportunity to capture market share. Their claims of improved driver support, backed by new AI-optimized testing procedures and expanded hardware configurations in their lab, suggest they’re taking this launch seriously.
Frequently Asked Questions
Q: What makes the Radeon 97 series different from previous AMD GPUs?
The Radeon 97 series represents a return to monolithic die design and features significant improvements in compute unit performance (40% faster per unit), enhanced ray tracing capabilities, and new AI features. It also includes improved media encoding capabilities and unrestricted encoding sessions.
Q: How does AMD’s new ray tracing technology compare to NVIDIA’s?
While it may not fully match NVIDIA’s capabilities, AMD has made substantial improvements with their third-generation ray tracing accelerators, including doubled throughput for ray testing and dedicated hardware for ray transformation. The actual performance difference will be clearer once independent reviews are published.
Q: What are the key differences between the 9770 and 9770 XT models?
The 9770 XT features higher clock speeds (near 3GHz), 8 additional compute units and RT accelerators, 16 more AI accelerators, and draws 84 watts more power. Both cards have 16GB of memory on a 256-bit bus, but the XT model offers significantly more performance for just $50 more.
Q: Will these GPUs support AI workloads?
Yes, but with limitations. The 16GB VRAM capacity restricts the size of AI models they can handle. However, they offer improved performance for supported models, with 2x better FP16 and 4x better INT8 performance compared to the previous generation.
Q: When will the Radeon 97 series be available for purchase?
AMD has announced wide availability starting March 6, 2025. Unlike previous launches, these cards have already been shipping to retailers for months, suggesting better initial stock availability than recent GPU launches from competitors.
Finn is an expert news reporter at DevX. He writes on what top experts are saying.
