Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more

The rumor was right about the headline numbers: NVIDIA’s GeForce RTX 5090 officially has a 512-bit memory interface and 28-Gbps GDDR7. NVIDIA later published those specifications for the RTX 5090; they were not confirmed when the March 2024 rumor first circulated.

Which graphics card has the rumored specifications?

The specifications belong to the GeForce RTX 5090, NVIDIA’s Blackwell-generation flagship. NVIDIA’s architecture whitepaper identifies the RTX 5090 as the first GeForce RTX 50 Series card to use the GB202 GPU. The original March 2024 reports attributed the 512-bit claim to leaker Kopite7kimi, rather than to NVIDIA. NVIDIA’s January 2025 announcement and Blackwell architecture whitepaper later established the production specification.

At the time, rumor coverage also reported disagreement over whether Blackwell gaming GPUs would keep a 384-bit maximum bus, and the bus widths of individual models were not known. Lowyat reproduced Kopite7kimi’s statement that the gaming flagship would have a 512-bit interface; that was a leaker’s claim, not an NVIDIA confirmation. TechSpot’s March 2024 report and Lowyat’s report provide the contemporaneous rumor context.

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.

RTX 5090 memory specifications

Specification GeForce RTX 5090
Memory capacity 32 GB
Memory type GDDR7
Memory interface 512-bit
Memory data rate 28 Gbps
Aggregate memory bandwidth 1,792 GB/s

NVIDIA’s 2025 whitepaper describes the full GB202 GPU as having a 512-bit memory interface with sixteen 32-bit memory controllers. The product specifications list 32 GB of GDDR7, a 28-Gbps data rate and 1,792 GB/s of bandwidth. Capacity, interface width, data rate and bandwidth describe different aspects of memory; 32 GB is the amount of memory, not its speed.

#1 Best Overall
ASUS TUF Gaming GeForce RTX™ 5080 16GB GDDR7 OC Edition Graphics Card
  • Powered by the NVIDIA Blackwell architecture and DLSS 4. System Requirements: Minimum 850W PSU with 16-pin 12V-2x6 (12VHPWR) connector required. Verify before purchasing.
  • Military-grade components deliver rock-solid power and longer lifespan for ultimate durability. Compatibility: 348mm (13.7") length, 3.6 slots, 4.3 lbs. Confirm case clearance and slot spacing. GPU bracket included.
  • Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
  • 3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans
  • Phase-change GPU thermal pad helps ensure optimal thermal performance and longevity, outlasting traditional thermal paste for graphics cards under heavy loads

How 28 Gbps becomes 1,792 GB/s

The 28-Gbps figure is a per-pin data rate, measured in gigabits per second. Multiply it by the 512-bit interface width to get 14,336 gigabits per second across the interface; divide by eight to convert bits to bytes. The result is 1,792 gigabytes per second of theoretical aggregate bandwidth.

That bandwidth figure does not mean every game or application will transfer data at that rate continuously. Real-world performance depends on more than memory bandwidth, so a 512-bit interface or a larger bandwidth number alone cannot establish how fast the card will run a particular game.

Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Support on Ko-Fi

How the RTX 5090 compares with other RTX 50 desktop cards

NVIDIA’s product comparison lists these memory specifications for four GeForce RTX 50 Series desktop models:

Special offer. See more information about Outbyte and uninstall instructions. Please review EULA and Privacy policy.
Model Memory Interface Data rate Bandwidth
GeForce RTX 5090 32 GB GDDR7 512-bit 28 Gbps 1,792 GB/s
GeForce RTX 5080 16 GB GDDR7 256-bit not stated in the cited comparison 960 GB/s
GeForce RTX 5070 Ti 16 GB GDDR7 256-bit not stated in the cited comparison 896 GB/s
GeForce RTX 5070 12 GB GDDR7 192-bit not stated in the cited comparison 672 GB/s

The listed capacity, interface and bandwidth figures come from NVIDIA’s GeForce RTX 50 Series comparison; it does not state data rates for the other three models in that comparison. These memory figures are not a ranking of gaming performance or value. Establishing either requires relevant performance tests and prices, neither of which follows from bus width alone.

What the specifications do—and do not—confirm

The original report was a rumor about a future flagship, and NVIDIA later confirmed the same headline memory interface and data rate for the RTX 5090. That match does not turn the earlier leak into an official specification retroactively. The numbers are useful for understanding the RTX 5090’s memory subsystem, but they are not a substitute for application-specific benchmarks or current price and availability information.

Quick Recap

Bestseller No. 1
ASUS TUF Gaming GeForce RTX™ 5080 16GB GDDR7 OC Edition Graphics Card
ASUS TUF Gaming GeForce RTX™ 5080 16GB GDDR7 OC Edition Graphics Card
3.6-slot design with massive fin array optimized for airflow from three Axial-tech fans; Auto-Extreme precision automated manufacturing helps ensure higher reliability
$1,831.31

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.