Search
HW News

NVIDIA Pulls RTX 5090, Shifts GB202 to RTX Pro [2026]

Owen Castellanos
Owen CastellanosHardware & Gaming Reporter
12 min read
NVIDIA Pulls RTX 5090, Shifts GB202 to RTX Pro [2026]

NVIDIA is reportedly walking away from its own flagship gaming card. Multiple hardware outlets, including VideoCardz, TechPowerUp, and TweakTown, reported on October 10 and 11, 2026 that the GeForce RTX 5090 is being phased out, with every remaining GB202 die redirected to the professional RTX PRO lineup. The claim traces back to hardware leakers rather than an NVIDIA press release, but the story has now been picked up by enough independent outlets that it has become the dominant hardware headline of the week.

NVIDIA has not issued a statement confirming the GeForce RTX 5090 is discontinued, and the company has given no end-of-production date. It also has not confirmed that future GB202 supply is reserved exclusively for RTX PRO products. Everything below that goes beyond those two confirmed facts is sourced to reports, leaker claims, and named outlets, and is labeled as such throughout this piece.

Google · Preferred Sources

Don't miss new tech stories on Google

Add TrendinTech once in the Google app and our stories appear in your news suggestions.

Add Now

What the Leakers Are Actually Claiming

The story originates with hardware leaker MEGAsizeGPU, who posts under the handle @Zed__Wang. According to reporting from TechPowerUp, the leaker wrote “No more GB202 for 5090, confirmed,” then followed up with a second post stating “all future GB202 supplies are RTX Pro exclusive.” Those two lines are the entire basis for the discontinuation claim, and they are the leaker’s own words, not an NVIDIA statement.

A second leaker, identified in VideoCardz’s report as hongxing2020, added a related claim: that NVIDIA had issued a factory notification stating the company would no longer sell any RTX 5090 models, including the RTX 5090 D variant built for the Chinese market. That detail remains unconfirmed and applies, if accurate, to a specific regional product line rather than the entire RTX 5090 family worldwide.

It is worth separating the two claims being made here, since outlets have sometimes blurred them together. The first is a supply-allocation claim: GB202 dies are being steered toward RTX PRO. The second is a production claim: the RTX 5090 itself is ending. The two are related but not identical, since NVIDIA could theoretically keep assembling RTX 5090 boards from existing GB202 inventory even while halting new die allocations, a scenario several outlets have floated without confirmation.

GB202: The Silicon at the Center of the Story

GB202 is the GPU die that powers the GeForce RTX 5090, NVIDIA’s current consumer flagship. It is also, according to the same reports, the die family NVIDIA uses for its RTX PRO workstation and data center cards. That shared lineage is precisely why the rumor has traction: if GB202 wafer output is finite and NVIDIA’s enterprise and AI customers are willing to pay substantially more per chip than gaming customers, reallocating supply away from the GeForce lineup is a financially rational move for the company to make, even if NVIDIA has not said it is doing so.

This is not the first time GB202 silicon economics have made news this year. GDDR6 and GDDR7 memory pricing has already been climbing industry-wide, and GPU die allocation decisions compound that pressure by squeezing the number of finished gaming cards that reach retail. When the chip powering your halo product is also the chip your highest-margin enterprise customers want, the gaming SKU is the one that tends to lose the allocation fight.

Why RTX PRO Demand Is Reportedly Winning the Allocation Fight

NVIDIA’s RTX PRO series targets workstation customers doing 3D rendering, scientific computing, and increasingly, local AI inference and fine-tuning work. Demand for AI-capable hardware has been one of the defining stories of the past two years, and NVIDIA’s data center and professional segments have consistently commanded higher per-unit margins than consumer gaming GPUs. If the leaked claims are accurate, redirecting GB202 output toward RTX PRO would let NVIDIA sell the same silicon at a higher price point to a customer base with less sensitivity to cost, at the expense of gaming availability.

None of this has been confirmed by NVIDIA, and the company has offered no explanation of wafer allocation, production volumes, or customer mix for GB202. The reasoning above reflects the general logic industry analysts have applied to similar allocation stories in the past, not a confirmed NVIDIA rationale for this specific case.

NVIDIA’s Silence Is Its Own Story

What makes this story unusual is how it has traveled. There is no earnings call comment, no investor note, and no product page change tied to it. Every outlet covering the claim, from Tom’s Hardware to Guru3D, frames the claim as unconfirmed and sourced to social media leakers. That has not stopped the story from spreading, in part because MEGAsizeGPU and other leakers in this space have accurately predicted NVIDIA product decisions before, lending their posts more weight than anonymous rumors typically carry.

Still, a track record of accurate past leaks is not the same as official confirmation, and readers should treat every specific detail here, including any timeline for when RTX 5090 stock might run dry at retail, as unverified until NVIDIA comments directly.

Timeline: How the Story Broke

The table below tracks how the claim moved across the hardware press in the span of about 24 hours, based on publication dates from each outlet’s own reporting.

OutletReport DateCore Claim Reported
TechPowerUp2026-10-10MEGAsizeGPU post claiming GB202 reallocation to RTX PRO
Hardware Busters2026-10-10Single leaker source, NVIDIA has not commented
VideoCardz2026-10-10Two leakers cited, GB202 steered away from RTX 5090
HotHardware2026-10-10GB202 dies redirected toward AI and workstation lines
TweakTown2026-10-10RTX 5090 described as nearing end of life
Guru3D2026-10-10NVIDIA reportedly stops supplying GB202 to RTX 5090 cards
Tom’s Hardware2026-10-10Production halt claim plus RTX 5080 24GB successor rumor
XenoSpectrum2026-10-10Labeled explicitly as rumor, no end date given

The clustering of nearly every report on the same single day is itself notable. It suggests the hardware press moved quickly on a leaker post rather than independently verifying the claim through separate supply-chain sources, which is common practice for this kind of story but also means the underlying evidence base is thinner than the volume of coverage might suggest.

Confirmed vs. Unconfirmed: Separating Fact From Rumor

Given how much of this story rests on social media posts, it is worth laying out exactly what counts as confirmed reporting versus what remains an open claim.

ClaimStatusBasis
GB202 powers the GeForce RTX 5090ConfirmedEstablished product fact
NVIDIA has announced RTX 5090 discontinuationNot confirmedNo NVIDIA statement exists
NVIDIA has confirmed GB202 is RTX PRO exclusive going forwardNot confirmedNo NVIDIA statement exists
GB202 silicon is being redirected away from RTX 5090UnconfirmedLeaker claim (MEGAsizeGPU)
RTX 5090 production has endedUnconfirmedLeaker claim, multiple outlets repeating it
RTX 5090 D (China variant) is affectedUnconfirmedLeaker claim (hongxing2020)
A replacement product, price, or launch date existsNot confirmedNo official details on record

Historical Context: NVIDIA Has Trimmed Flagship Cards Before

NVIDIA reshuffling its product stack mid-generation is not without precedent. The company has a long history of adjusting SKUs, memory configurations, and allocation priorities within a single GPU generation when market conditions shift, whether that shift is driven by mining demand, AI demand, or component shortages. What makes the current reported situation distinctive is the target: a current-generation flagship gaming card being squeezed not by a shortage of raw materials, but reportedly by internal competition from NVIDIA’s own professional product line for the exact same die.

That dynamic mirrors a broader pattern across the GPU and CPU industry in 2026, where manufacturers increasingly treat consumer and enterprise silicon as a single allocatable pool rather than separate production tracks. Chipmakers across both the GPU and CPU space have leaned harder into enterprise and AI-tied product lines this year, often at the expense of the consumer upgrade cycle that used to anchor annual product refreshes.

Market Impact: What Happens to RTX 5090 Pricing and Availability

If GB202 allocation to the RTX 5090 is genuinely winding down, the most immediate practical consequence for shoppers is a supply drought rather than an instant disappearance from shelves. Tom’s Hardware’s reporting frames this as an “impending supply drought” that would be expected to push prices upward as existing retail and board-partner inventory thins out, well before any card physically vanishes from store listings. That is a standard pattern whenever a GPU nears the end of its production run: prices rarely drop, and street prices for remaining units tend to climb as the card becomes harder to restock.

For anyone currently comparing high-end GPU options, it is worth noting this uncertainty sits alongside an already turbulent memory and component pricing environment. Flagship and upper-midrange GPU pricing has been a moving target across both NVIDIA and AMD’s lineups this year, and a reported RTX 5090 supply pullback would only add to that volatility rather than ease it.

What a RTX 5080 24GB Successor Rumor Means

Several of the same reports, including Tom’s Hardware and an Italian outlet cited across the coverage, mention chatter about a 24GB RTX 5080 variant potentially becoming the new top-tier GeForce gaming option if the RTX 5090 truly exits the lineup. NVIDIA has given no official details on any such product, including no confirmed memory configuration, price, or launch date. Readers should treat this strictly as a secondary rumor riding alongside the main discontinuation claim, not as a confirmed roadmap item.

Why Gamers Should Be Skeptical, Not Panicked

It is easy to read headlines like “NVIDIA reportedly discontinuing RTX 5090” and assume the card is already gone. It is not. Existing RTX 5090 units remain available through normal retail and board-partner channels as of this writing, and NVIDIA has made no statement changing that. The practical reality for most buyers right now is unchanged: the card you can order today is the same card you could order last week. What has changed is the confidence level around how long that will remain true, and that confidence level rests entirely on unverified leaker claims rather than NVIDIA’s own roadmap communication.

This is also a useful moment to separate two different kinds of hardware uncertainty that often get conflated in headlines: a confirmed end-of-life announcement from the manufacturer, and a leaked internal allocation decision that has not been verified by the manufacturer. The RTX 5090 situation, as reported right now, is firmly the second kind.

Competitive Landscape: Where AMD and Intel Fit In

A reported RTX 5090 supply pullback lands at an interesting moment competitively. AMD’s current flagship consumer GPU sits a tier below the RTX 5090 in NVIDIA’s stack, and the two companies have been trading blows on pricing and memory configuration rather than raw halo-tier performance for most of 2026. If NVIDIA genuinely steps back from defending the absolute top of the gaming GPU market in favor of RTX PRO margins, that creates an opening, at least temporarily, for AMD to make a stronger pitch at the enthusiast segment, even without matching NVIDIA chip-for-chip at the very top end.

Intel’s discrete GPU ambitions remain focused on the mainstream and midrange segments rather than flagship competition, so this particular story has limited direct bearing on Intel’s graphics roadmap. The more relevant competitive read-through is on the CPU and platform side, where component allocation decisions throughout the PC hardware supply chain have been increasingly shaped by AI-driven enterprise demand rather than traditional consumer upgrade cycles.

Expert and Industry Voices

The claims driving this story originate almost entirely from hardware leakers rather than NVIDIA executives or named analysts, and it is important to represent their statements accurately rather than paraphrase them as confirmed fact. According to TechPowerUp’s reporting, leaker MEGAsizeGPU, posting as @Zed__Wang, wrote: “No more GB202 for 5090, confirmed.” In a follow-up post reported by the same outlet, the leaker added: “all future GB202 supplies are RTX Pro exclusive.”

A second leaker, cited in VideoCardz’s report, hongxing2020, is said to have shared a factory notification stating: “We have received a notification from the factory: NVIDIA will no longer sell any RTX 5090 models, including the D variants.” Both leakers’ claims remain unverified by NVIDIA, and neither statement should be read as an NVIDIA press release or official product announcement.

Predictions: What Happens From Here

Given the current state of the reporting, a few outcomes look more likely than others over the next several weeks, though all remain speculative until NVIDIA comments.

  • NVIDIA stays silent on the specific allocation claim for an extended period, neither confirming nor denying it directly, which is consistent with how the company has handled similar leaker stories in the past.
  • Retail RTX 5090 pricing drifts upward in the coming weeks if board partners confirm thinning allocations, even without an official discontinuation notice.
  • Rumors around a 24GB RTX 5080 variant intensify if RTX 5090 supply visibly tightens, giving NVIDIA a lower-cost flagship narrative to lean on.
  • AMD uses any visible RTX 5090 supply gap to push its own high-end positioning and pricing more aggressively in marketing.
  • Additional leakers or supply-chain sources either corroborate or contradict MEGAsizeGPU’s original claim, since a story this widely reported tends to attract competing tips within days.

What Buyers Should Actually Do Right Now

For anyone currently shopping in the enthusiast GPU tier, the practical advice is straightforward: buy based on currently confirmed specs and current pricing, not on a rumor about future scarcity. If the reported allocation shift is real and RTX 5090 supply does tighten, prices for existing inventory are more likely to rise than fall, so there is little downside to purchasing a card today if you already need one. If you can wait, it is reasonable to watch for NVIDIA’s response, or lack of one, over the coming weeks before making a final call on a flagship-tier purchase, particularly since broader hardware spending patterns have already been unusually soft across the industry this year, which could blunt how aggressively retailers mark up a thinning RTX 5090 supply even if the rumor proves accurate.

It is also worth remembering that board partners, not just NVIDIA directly, control a meaningful share of retail inventory and pricing decisions. Even in a scenario where NVIDIA stops shipping new GB202 dies for RTX 5090 boards, partners with existing chip stock could continue assembling and selling cards for some additional window of time, which is one reason outlets have avoided naming a specific end date for when RTX 5090 cards might actually disappear from shelves.

The Bigger Picture: Gaming GPUs in an AI-First Chip Market

Whether or not this specific claim about the RTX 5090 turns out to be accurate, the underlying tension it describes is real and has been building for the better part of two years. GPU manufacturers are increasingly allocating their most advanced silicon toward AI and professional workloads, where per-unit margins dwarf what a gaming card can command at retail. That same reallocation pressure has already shown up on the CPU side of the industry this year, as manufacturers plan future platforms with an eye toward AI workloads as much as traditional desktop performance.

If NVIDIA genuinely is willing to sunset its own current-generation gaming flagship to feed RTX PRO demand, it would mark one of the clearest signals yet that the AI boom is reshaping even the top of the consumer hardware stack, not just data center roadmaps. That is the real story worth watching here, independent of whether this particular leak proves accurate in every detail.

Frequently Asked Questions

Has NVIDIA officially confirmed the RTX 5090 is discontinued?
No. NVIDIA has not issued any statement confirming discontinuation and has given no end-of-production date. The claim originates from hardware leakers and has been reported by multiple outlets as unconfirmed.

What is GB202 and why does it matter?
GB202 is the GPU die that powers the GeForce RTX 5090. It is reportedly the same die family used in NVIDIA’s professional RTX PRO cards, which is why leakers claim NVIDIA could redirect supply from gaming to enterprise customers.

Who first reported this claim?
Hardware leaker MEGAsizeGPU (@Zed__Wang) posted the original claim on X, followed by a second leaker, hongxing2020, who cited an alleged factory notification. Outlets including TechPowerUp, VideoCardz, and TweakTown then reported on those posts.

Can I still buy an RTX 5090 today?
As of this writing, RTX 5090 cards remain available through normal retail and board-partner channels. Nothing in NVIDIA’s official communications has changed current availability.

Does this affect the RTX 5090 D sold in China?
One leaker claim suggests it does, stating NVIDIA would stop selling RTX 5090 models including the D variant. This detail is unconfirmed and specific to that single source.

Will prices go up if this is true?
If GB202 allocation to RTX 5090 genuinely tightens, reporting from outlets like Tom’s Hardware suggests a supply drought and resulting price increases are the expected pattern, though no specific pricing figures have been confirmed.

Is there a confirmed replacement for the RTX 5090?
No. Rumors mention a possible 24GB RTX 5080 variant, but NVIDIA has given no official details on any replacement product, price, or launch date.

Why would NVIDIA prioritize RTX PRO over its own flagship gaming card?
Reports suggest RTX PRO and AI-focused customers generate higher per-unit margins than consumer gaming sales, making a reallocation of limited GB202 supply financially appealing if true, though NVIDIA has not confirmed this reasoning.

Related Coverage

Owen Castellanos

Owen Castellanos

Hardware & Gaming Reporter

Owen Castellanos is the Hardware & Gaming Reporter at TrendinTech, where he tests graphics cards, processors, consoles and peripherals and follows the games industry, from the major publishers to independent studios. He previously wrote reviews and benchmarks for Tom's Hardware and contributed features to PC Gamer, covering everything from laptop testing to the economics of game development. Owen holds a Bachelor of Science in Electrical Engineering from the University of Texas at Austin and attends CES in Las Vegas, Computex in Taipei and the Game Developers Conference in San Francisco each year to meet chipmakers and developers. He writes for readers who want to know whether a product is worth the money, and he is not shy about saying when it is not.

All stories by Owen Castellanos (287)

Related Articles