Newsletter IconFacebook IconX IconThreads IconInstagram IconYouTube IconPinterest Icon
Giveaway: Win an NZXT H6 RGB+ Case, Kraken Elite AIO, RGB Fans and 1200W PSU

NVIDIA is reportedly preparing the GeForce RTX 5090 SE to sit between the RTX 5080 and 5090

NVIDIA's rumored RTX 5090 SE could bridge the gap between the RTX 5080 and RTX 5090, but leaked memory specs and pricing don't add up yet.

NVIDIA is reportedly preparing the GeForce RTX 5090 SE to sit between the RTX 5080 and 5090
CommentsFacebook IconX IconReddit Icon
Tech Reporter
Published
2 minutes & 45 seconds read time
TL;DR: Leaked reports claim NVIDIA may launch an RTX 5090 SE positioned between the RTX 5080 and 5090, using a cut-down GB202 die with 14,080 CUDA cores, ~500W power, and a disputed 32GB GDDR7 on a 384-bit bus; pricing and memory details remain questionable.
Voice: Hassam Nasir
0:00 / 3:58
Use left and right arrow keys to seek audio.

NVIDIA is reportedly working on a new GPU called the GeForce RTX 5090 SE, according to a report from Russian outlet GameGPU. The card would supposedly slot between the RTX 5080 and RTX 5090, possibly replacing an RTX 5080 Ti that never made it to market.

The leaked specs point to a cut-down GB202 die, with 110 of the 192 streaming multiprocessors enabled. That works out to 14,080 CUDA cores, roughly 31% more than the RTX 5080's 10,752 cores, but still well short of the RTX 5090's 21,760. Rumors also point to a 500W power draw, down from the 5090's 575W, but a big step up from the 5080's 360W. A 1,000W power supply would likely be needed either way.

Now we get to the confusing part. The leak claims 32GB of GDDR7 on a 384-bit bus, the same capacity as the RTX 5090 despite the narrower memory interface. That math doesn't really work with NVIDIA's standard memory configurations. A 384-bit bus paired with 2GB modules gets you 24GB, and 3GB modules would push it to 36GB. Hitting exactly 32GB would apparently require mixing chip densities, something NVIDIA has avoided in the past due to signal integrity and validation headaches, not to mention the GTX 970 controversy that mixed-pool memory caused years ago.

Image By Techpowerup
Image By Techpowerup

It's also hard to ignore the optics here. NVIDIA is reportedly ready to cram 32GB onto a GPU most people will never afford, while the budget segment is stuck fighting over scraps. The RTX 5050 9GB, which would have bumped entry-level VRAM from 8GB to 9GB, was reportedly canceled just weeks ago, with NVIDIA leaning on a reissued RTX 3060 12GB to cover that segment instead.

NVIDIA is reportedly preparing the GeForce RTX 5090 SE to sit between the RTX 5080 and 5090 2

So the story right now is a company allegedly finding enough GDDR7 to build a near-5090 for enthusiasts with deep pockets, while gamers on a budget get handed a five-year-old Ampere card as the "solution." Memory supply is tight industry-wide, sure, but where NVIDIA chooses to point that limited supply says a lot about who it thinks its customers actually are.

The rumored price sits at $1,500, which also looks shaky given how GPU pricing has been trending lately. The RTX 5090 already carries an MSRP of $1,999, and street prices have been running well above that thanks to ongoing memory shortages. It's hard to see NVIDIA slotting in a nearly-as-loaded 32GB card for $500 less while memory costs are what's driving prices up in the first place.

NVIDIA is reportedly preparing the GeForce RTX 5090 SE to sit between the RTX 5080 and 5090 4

Frequently Asked Questions

TweakBot answers common questions about this news using TweakTown's own coverage from this page and related content from our archive. Tap a question to reveal the answer, or type your own below.

Question #1

What power supply recommendations and connector requirements should builders consider for a card with the rumored 500W draw?

You should plan on roughly a 1,000W power supply to support a card rumored to draw about 500W, per the primary report. Also ensure you have proper high-current GPU power connectors and wiring: reports note 12VHPWR connectors have melted at high temperature ("GeForce RTX 5090 power connector spotted at 150+ Celsius, prone to melting") and a prototype showed a 4 x 16-pin power connector layout for the RTX 5090 ("NVIDIA's prototype GeForce RTX 5090 leaked: rocks a crazy 4 x 16-pin power connector setup").
Answered
Question #2

Could NVIDIA realistically ship a 32GB GDDR7 configuration on a 384-bit bus without mixed-density memory modules, based on past TweakTown coverage?

No. The article explains a 384-bit bus with standard 2GB modules yields 24GB and 3GB modules yield 36GB, so 32GB would require mixing chip densities, something NVIDIA has avoided due to signal integrity, validation headaches, and the GTX 970 mixed-pool controversy. Thus shipping 32GB on a 384-bit bus without mixed-density modules is not realistic according to the coverage.
Answered
Question #3

How might the alleged RTX 5090 SE $1,500 price interact with current RTX 5090 street prices and availability in retail?

If NVIDIA listed the RTX 5090 SE at $1,500 it would likely conflict with current RTX 5090 street prices, which are running well above the RTX 5090 MSRP of $1,999 due to memory shortages. The primary article argues it is hard to see NVIDIA offering a near-5090 32GB card for $500 less while memory costs push prices up, and it expects actual retail pricing to land higher than the official price following this generation's pattern.
Answered
Question #4

If the 5090 SE uses a cut-down GB202 die, what implications would that have for cooling solutions and GPU cooler designs compared with current 5090/5080 models?

A cut-down GB202 die with 110 SMs and a rumored 500W draw would still demand far more cooling capacity than the RTX 5080’s 360W and sit closer to the RTX 5090’s 575W, so GPU coolers will need larger heatsinks, higher airflow or denser fins, stronger fans or more aggressive fan curves, and beefed-up VRM and memory cooling to manage the extra heat. That higher thermal budget also reinforces the likelihood of users favoring robust power and cooling solutions such as custom or AIO water blocks; a custom water block on the RTX 5090 showed large temperature drops in practice (see Related 1: NVIDIA GeForce RTX 5090 FE with a custom water block).
Answered

Have a question not listed here? Ask below and TweakBot will answer it.

If the RTX 5090 SE is real, expect actual retail pricing to land higher than whatever NVIDIA puts on the box, following the same pattern as every other GPU launch this generation. For now, treat the specs with some skepticism until more sourcing shows up.

Photo of the GIGABYTE GeForce RTX 5090 Gaming OC Graphics Card

Best Deals: GIGABYTE GeForce RTX 5090 Gaming OC Graphics Card

Prices last scanned 1 hour and 32 minutes ago

* Prices may be inaccurate. As an Amazon Associate, we earn from qualifying purchases. We earn affiliate commission from any Newegg or PCCG sales.

News Source:en.gamegpu.com

Comments

Tech Reporter

Email IconX IconLinkedIn Icon

Hassam is a veteran tech journalist and editor with over eight years of experience embedded in the consumer electronics industry. His obsession with hardware began with childhood experiments involving semiconductors, a curiosity that evolved into a career dedicated to deconstructing the complex silicon that powers our world. From benchmarking PC internals to stress-testing flagship CPUs and GPUs, Hassam specializes in translating high-level engineering into deep, unbiased insights for the enthusiast community.

Stay Updated

Follow TweakTown for breaking tech news, reviews, and daily updates.

Add TweakTown as a preferred source on GoogleFind TweakTown on Apple News
Newsletter Subscription