Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
The 2024 RDNA 4 memory leak was broadly right about AMD’s bus-width and Infinity Cache strategy, but it was not an exact specification sheet. AMD has since launched the Radeon RX 9000 series, confirming the 256-bit/64MB Navi 48-class configuration, the 192-bit/48MB cut-down design, and the 128-bit/32MB Navi 44-class tier. Some details changed: the rumored 19 Gbps configuration became an 18 Gbps Radeon RX 9070 GRE, and the leaked RX 8000 branding gave way to Radeon RX 9000.
What the original leak claimed
The report, published while RDNA 4 was still unreleased, summarized social-media claims attributed to @all_the_watts, with links to earlier claims from @Kepler_L2. These were rumors, not AMD documentation, and they did not establish final product names or retail-card mappings.
| Leaked GPU configuration | Bus | Reported memory speed | Theoretical bandwidth | Reported Infinity Cache |
|---|---|---|---|---|
| Navi 48 high-end | 256-bit | 20 Gbps | 640 GB/s | 64MB |
| Navi 48 lower-speed variant | 256-bit | 18 Gbps | 576 GB/s | 64MB |
| Navi 48 cut-down variant | 192-bit | 19 Gbps | 456 GB/s | 48MB |
| Navi 44 | 128-bit | 18 Gbps | 288 GB/s | 32MB |
The leak therefore described a segmented memory hierarchy: wider interfaces and larger caches for Navi 48, followed by a narrower 128-bit Navi 44 design for lower-tier products. It did not prove performance, launch dates, die sizes, pricing, or whether every configuration would reach stores.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesHow the bandwidth figures are calculated
Raw memory bandwidth is calculated with a simple formula:
#1 Best Overall
- Axial-tech fans now feature a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- 2.5-slot design allows for greater build compatibility while maintaining cooling performance
- 0dB technology lets you enjoy light gaming in relative silence
- Dual BIOS switch lets you toggle between Quiet and Performance BIOS profiles
- Dual ball fan bearings last up to twice as long as sleeve bearing designs
Bandwidth = memory-transfer rate × memory-bus width ÷ 8The division by eight converts bits into bytes. Applying the formula to the leaked figures gives:
20 Gbps × 256 ÷ 8 = 640 GB/s18 Gbps × 256 ÷ 8 = 576 GB/s19 Gbps × 192 ÷ 8 = 456 GB/s18 Gbps × 128 ÷ 8 = 288 GB/s
These are raw theoretical external-memory bandwidth figures. They do not translate directly into frame rates. Actual results depend on the GPU’s compute resources, memory compression, cache hit rate, workload, resolution, and software.
What AMD eventually shipped
AMD’s official specifications and ROCm GPU documentation confirm the following shipping configurations:
Rank #2
- Axial-tech fans now feature a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- Phase-change GPU thermal pad helps ensure optimal heat transfer, lowering GPU temperatures for enhanced performance and reliability
- 2.5-slot design allows for greater build compatibility while maintaining cooling performance
- Dual-ball fan bearings last up to twice as long as standard conventional sleeve bearings designs
- 0dB technology lets you enjoy light gaming in relative silence
| Card | GPU family | Compute units | VRAM | Bus | Memory speed | Raw bandwidth | Infinity Cache |
|---|---|---|---|---|---|---|---|
| Radeon RX 9070 XT | Navi 48-class | 64 | 16GB GDDR6 | 256-bit | Up to 20 Gbps | Up to 640 GB/s | 64MB |
| Radeon RX 9070 | Navi 48-class | 56 | 16GB GDDR6 | 256-bit | Up to 20 Gbps | Up to 640 GB/s | 64MB |
| Radeon RX 9070 GRE | Navi 48-class | 48 | 12GB GDDR6 | 192-bit | Up to 18 Gbps | Up to 432 GB/s | 48MB |
| Radeon RX 9060 XT 16GB | Navi 44-class | 32 | 16GB GDDR6 | 128-bit | Up to 20 Gbps | Up to 320 GB/s | 32MB |
| Radeon RX 9060 XT 8GB | Navi 44-class | 32 | 8GB GDDR6 | 128-bit | Up to 20 Gbps | Up to 320 GB/s | 32MB |
| Radeon RX 9060 | Navi 44 attribution should be treated cautiously | Not listed here | Not listed here | 128-bit | Up to 18 Gbps | Up to 288 GB/s | Not established here |
AMD’s product pages confirm the RX 9070 XT’s 16GB, 256-bit interface, 20 Gbps GDDR6, 640 GB/s bandwidth and 64MB cache. The RX 9060 XT has a 128-bit interface, 20 Gbps GDDR6, 320 GB/s bandwidth and 32MB cache. AMD separately lists the RX 9060 at up to 18 Gbps and 288 GB/s, but the available specification does not establish every cache and compute detail.
Leak versus reality
| Leak expectation | Shipping result | Assessment |
|---|---|---|
| Navi 48, 256-bit, 20 Gbps, 64MB | RX 9070 XT and RX 9070 use 256-bit memory, up to 20 Gbps and 64MB cache | Essentially confirmed |
| Navi 48 cut-down model, 192-bit, 19 Gbps, 48MB | RX 9070 GRE uses 192-bit memory, 18 Gbps, 432 GB/s and 48MB cache | Directionally correct; memory speed differed |
| Navi 44, 128-bit, 18 Gbps, 32MB | RX 9060 broadly matches the 18 Gbps/288 GB/s bandwidth class | Broadly correct, but exact SKU attribution should be qualified |
| Navi 44 with faster memory | RX 9060 XT uses 20 Gbps memory, 320 GB/s and 32MB cache | Correct architecture-level direction, not the exact leaked speed |
The leak got the broad architecture right: AMD retained the expected cache tiers and used bus widths that scale with the product segment. It did not predict every final SKU exactly. Reports also used the then-expected Radeon RX 8000 branding, while AMD ultimately launched the desktop family as Radeon RX 9000.
Pre-launch coverage from Tom’s Hardware appropriately treated higher-tier specifications as more reliable than lower-tier details, where Navi 44 and product segmentation remained uncertain.
Rank #3
- System Compatibility Note: 2.5‑slot card measuring 303 mm (L) x 131 mm (W) x 45 mm (H); requires a single 8‑pin power connector and a recommended 550W power supply. Please verify chassis clearance and power supply capacity before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- AMD RDNA 3 Architecture with AI & Ray Tracing Acceleration: Powered by 32 RDNA 3 Compute Units featuring 3rd Gen Ray Tracing Accelerators and 2nd Gen AI Accelerators, delivering lifelike lighting, shadows, and superior machine learning performance for enhanced gaming and content creation.
- Powerful 1080p & 1440p Gaming Engine: Features a max boost clock of up to 2695 MHz, a game clock of 2280 MHz, and 2048 stream processors, ensuring outstanding frame rates in the latest titles.
- 8GB High‑Speed GDDR6 Memory: Equipped with 8GB of GDDR6 memory on a 128‑bit interface running at 18 Gbps, delivering up to 288 GB/s bandwidth for high‑resolution textures and demanding game workloads.
Why Infinity Cache matters
Infinity Cache is an on-die cache designed to keep frequently reused data close to the GPU. When requested data is found in the cache, the GPU can avoid some external GDDR6 traffic. That can reduce pressure on the memory bus, improve effective memory behavior for cache-friendly workloads, and potentially improve performance per watt compared with achieving the same result solely through a wider interface.
Free tools Windows power users keep installed
One-click scans. No signup required.
The key limitation is locality. A game that repeatedly reuses textures, buffers, or other data can benefit more than a workload that continuously streams large, changing datasets. Resolution, asset size, compression, engine behavior, ray-tracing data structures, and cache hit rates all matter.
Infinity Cache is therefore not a fixed bandwidth multiplier. A 64MB cache does not turn 640 GB/s into a guaranteed larger number, and it is not equivalent to 64MB of extra VRAM. AMD lists 64MB on the RX 9070 XT and RX 9070, 48MB on the RX 9070 GRE, and 32MB on the RX 9060 XT.
Rank #4
- System Compatibility Note: 2.5-slot card, 290x123x51mm, two 8-pin power, recommended 700W PSU. Verify chassis clearance before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- AMD RDNA 4 Architecture: RX 9070 GPU with 56 CUs, 3584 stream processors, 3rd gen RT and 2nd gen AI accelerators – built for 1440p/4K gaming.
- Factory Overclocked Performance: Boost clock up to 2520 MHz, game clock 2070 MHz – delivers smooth, high-framerate gaming out of the box.
- 16GB GDDR6 on 256-Bit Bus: High-speed 20 Gbps memory provides exceptional bandwidth for 4K textures, ray tracing, and demanding workloads.
Does a 128-bit bus make Navi 44 too slow?
Not by itself. The RX 9060 XT combines a 128-bit GDDR6 interface with up to 20 Gbps memory, 320 GB/s of raw bandwidth and 32MB of Infinity Cache. That is half the external bandwidth of an RX 9070-class card, but the RX 9060 XT is also a smaller, lower-power GPU with fewer compute resources and a different performance target.
The meaningful question is whether the memory subsystem is balanced with the GPU’s shader throughput and intended resolution. At 1080p, CPU limits and rendering throughput may matter more than bus width. At 1440p, the balance between 16GB of capacity, 320 GB/s of bandwidth and the GPU’s compute performance becomes more important. At 4K, heavy ray tracing, large texture packs and modded games can place greater pressure on both throughput and capacity.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Bus width alone is a poor buying metric. A narrower interface can be reasonable when paired with faster memory, cache, compression and an appropriately sized GPU. It can still become a limitation in workloads with poor cache locality or sustained high-resolution streaming.
Best Value
- Powered by Radeon RX 9070 XT
- WINDFORCE Cooling System
- Hawk Fan
- Server-grade Thermal Conductive Gel
- RGB Lighting
Bandwidth is not VRAM capacity
These specifications describe different problems:
- Bandwidth is how quickly data can theoretically move between the GPU and external memory.
- Capacity is how much data can remain resident before assets must be streamed, evicted or reloaded.
- Infinity Cache is a smaller, faster on-die store that can reduce external-memory traffic for reusable data.
An 8GB RX 9060 XT can have the same 320 GB/s bandwidth and 32MB cache as the 16GB version, yet face capacity limits sooner in games with high-resolution textures, mods or demanding future assets. Conversely, 16GB does not guarantee higher frame rates if the GPU’s compute performance or raw bandwidth is the limiting factor.
For buyers who expect to keep a card for several years, the 16GB RX 9060 XT is the more flexible configuration when the price difference is reasonable. The 8GB model should be evaluated against its actual price and the games being played, rather than its identical bandwidth figure.
Why AMD continued with GDDR6
The final RX 9000 desktop cards use GDDR6 rather than GDDR7. GDDR6 is a mature and comparatively economical platform, but it offers lower pin-level transfer rates than contemporary GDDR7 designs. AMD’s approach places more emphasis on bus sizing, Infinity Cache, memory compression and overall architectural efficiency.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →That is a trade-off, not automatically a defect. GDDR6 can help control platform cost and supply complexity, while the cache can reduce how often the GPU needs to reach external memory. The downside is lower raw bandwidth than a similarly wide GDDR7 interface and greater importance placed on choosing enough VRAM capacity for the intended workload.
What the leak did not establish
- It did not identify the final retail card associated with every configuration.
- It did not prove that every rumored speed would ship.
- It did not establish performance, cache hit rates or effective bandwidth.
- It did not prove launch dates, final pricing, die sizes or board-partner designs.
- It did not resolve early uncertainty over whether lower-tier products would use cut-down Navi 48 or Navi 44.
AMD’s official specifications are the appropriate source of truth for released cards. For example, AMD announced the RX 9070 at a launch-price signal of $549 on February 28, 2025, but that is not a current 2026 street-price claim; buyers should check live pricing and availability separately.
What this means for buyers
- RX 9070 XT and RX 9070: The best fit among these configurations for buyers prioritizing 1440p or 4K performance, 16GB VRAM, a 256-bit interface and 64MB Infinity Cache.
- RX 9070 GRE: A distinct 12GB/192-bit compromise with 432 GB/s bandwidth and 48MB cache. It offers a stronger memory subsystem than Navi 44-class cards but less capacity than the 16GB RX 9070 models.
- RX 9060 XT 16GB: The more durable Navi 44-class choice for texture-heavy games, mods, higher resolutions and longer ownership when priced sensibly.
- RX 9060 XT 8GB: Its identical bandwidth and cache do not eliminate the capacity disadvantage. Buy it only when the price and workload make 8GB acceptable.
- RX 9060: Its up-to-18 Gbps memory and 288 GB/s bandwidth place it in the entry-level configuration discussed by the leak, but undisclosed specifications should not be filled in by inference.
The original leak was a useful architectural preview, not a buying guide. Its strongest success was anticipating the relationship between bus width and Infinity Cache. Its weakest point was the assumption that rumored speeds, names and product mappings would transfer one-to-one into AMD’s final lineup.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
