AMD is making an ambitious move with the next evolution of its graphics-enhancement ecosystem, unveiling a major update that forgoes the usual numbered naming convention in favor of a more enigmatic and distinctive identity. This new suite of visual technologies, collectively branded “Redstone,” signals AMD’s strongest attempt yet to position its upscaling and rendering tools as true competitors to Nvidia’s dominant DLSS (Deep Learning Super Sampling) system. Although the platform’s most cutting-edge capabilities are designed to shine brightest on AMD’s newest generation of GPUs, Redstone isn’t just something PC gamers should care about. The innovations within this package are poised to reshape how visual processing is handled across consoles, handheld platforms, cloud gaming services, and even future hardware generations.
Why Redstone matters
One of the primary reasons modern video games are able to achieve impressive clarity at demanding resolutions like 4K—despite ever-increasing graphical complexity—is the growing reliance on AI-supported image reconstruction. Instead of relying purely on brute-force rendering power, today’s techniques generate frames at a lower internal resolution and use machine-learning models to rebuild them into crisp, high-quality output. This approach preserves performance while delivering visuals that often look indistinguishable from native 4K.
While this sounds simple in theory, the actual pipeline involves multiple layers of real-time software tools: data-driven reconstruction algorithms, motion vector analysis, temporal anti-aliasing, noise filtering, and finely tuned sharpening passes. “Redstone” is AMD’s effort to bring all these components under a single, unified umbrella—like a digital workshop where each tool can function independently or work collaboratively depending on the needs of a particular game engine.
What does Redstone actually do?
Though upscaling forms the foundation of Redstone, AMD’s vision extends far beyond reconstructing resolution. Many gamers, especially competitive players, demand not only visual fidelity but also extremely high frame rates. To achieve this, AMD is rolling out a new machine learning-driven frame generation system, which leverages advanced AI models to create entirely new in-between frames. These inserted frames artificially increase the perceived smoothness of gameplay, allowing titles to reach frame rates once impossible on current hardware.
AMD’s older FSR versions—like FSR 3.1—offered a rudimentary form of frame generation, but Redstone introduces a far more refined implementation. With enhanced motion prediction models, improved optical flow algorithms, and a redesigned latency pipeline, Redstone aims to deliver cleaner motion with fewer visual distortions. AMD demonstrated this by showing a side-by-side comparison in F1 24, where the previous technology produced noticeable flickering in shadows and high-contrast areas, while Redstone produced stable, consistent lighting transitions.
Nvidia’s DLSS 4 deserves credit for popularizing multi-frame generation, but its rollout was controversial. Some PC gamers felt the company was leaning too heavily on software techniques to compensate for minimal generational improvements between RTX GPU families. Meanwhile, Intel’s XeSS 2—also featuring frame interpolation—is poised to be far more hardware-agnostic, potentially giving Intel a unique advantage as the only vendor with a cross-platform frame generation solution.
The challenge of benchmarking frame generation
The introduction of these techniques complicates the process of evaluating gaming performance. Traditional metrics like “average fps” don’t tell the whole story when half the frames are artificially created. Nevertheless, AMD made bold claims that Redstone can elevate performance dramatically—such as pushing Call of Duty: Black Ops 7 from a sluggish 23 fps to an ultra-smooth 109 fps when using performance-mode upscaling. However, AMD stresses that users should already be hitting around 60 fps after upscaling before enabling frame generation. Otherwise, the AI interpolation process may produce ghosting, motion blur, jitter, or visual artifacts.
Frame generation is therefore not a magic button that fixes unoptimized games—it’s a performance amplifier intended to push already-playable titles well beyond their normal limits.
New ray-tracing enhancements powered by AI
Redstone isn’t just about upscaling and frame generation. AMD is also introducing Radiance Caching, a feature that uses AI-assisted algorithms to speed up ray-traced lighting calculations by predicting how light behaves in complex environments. Ray tracing produces more realistic shadows, reflections, and global illumination, but it is notoriously computationally expensive.
Another feature, Ray Regeneration, works to restore fine details that might be lost during ray-tracing passes. Together, these tools could allow games to offer more visually rich environments without tanking performance. However, enabling these features requires manual configuration inside the AMD Software control panel on a per-title basis, and individual games must be coded to support them.
Game support and future adoption
Currently, just over 200 PC games support some form of Redstone or FSR 4 integration, including large titles like Warhammer 40K: Darktide and Call of Duty: Black Ops 7. This puts Redstone in a competitive position relative to Nvidia’s DLSS 4 lineup, which includes around 175 supported titles as of late 2024. AMD says many more games will adopt the technology in 2026, though this ultimately depends on how aggressively AMD can lobby developers and engine makers to integrate the tools.
Why console players should pay close attention
Although Redstone’s most powerful features require RDNA 4 hardware—specifically the Radeon RX 9000-series GPUs like the RX 9070 and RX 9070 XT—the implications for console gaming are significant. Sony collaborated with AMD on FSR 4, released earlier this year, and leaks suggest Sony’s updated PSSR (PlayStation Spectral Super Resolution) pipeline will incorporate FSR-derived AI models to enhance performance on the upcoming PS5 Pro.
Because the existing PS5 is built on an older RDNA 2 foundation, integrating FSR 4 natively is a challenge. However, Redstone-level technology could be a defining feature of the rumored PlayStation 6, especially as next-gen consoles move toward hybrid AI-driven rendering pipelines. Upscaling technology is especially valuable for lower-power devices like handhelds, which could benefit dramatically from RDNA 4’s efficiency gains.
The bigger picture: software is redefining performance
AMD’s Redstone initiative marks more than just another update to FSR—it represents a broader shift in how the gaming industry views graphical performance. Historically, gaming hardware races were defined almost entirely by raw GPU power, transistor count, clock speed, and memory bandwidth. But now, the battle is increasingly fought on the software front, where AI-driven tools can multiply the effective output of existing hardware.
In the coming years, it’s likely that software-based enhancement will become just as important as physical silicon. Whether you’re gaming on a high-end PC, a living-room console, a cloud-streamed service, or a handheld device, the next generation of performance will rely heavily on reconstruction, interpolation, predictive lighting models, and intelligent upscaling—precisely the areas where AMD’s Redstone is staking its claim.