Quick Answer: GPU prices are spiking hard in August 2026. Nvidia just told some of its biggest data-center customers that AI server prices will rise by more than 15% on systems shipped in early 2027, Bloomberg reported on August 22. That’s layered on top of an existing surge in consumer graphics card prices: the RTX 5090 now sells for a median of around $4,700 — roughly 135% above its $1,999 launch MSRP — while every other card in the RTX 50 lineup is also selling well above its official price. The root cause is a global memory (DRAM and GDDR7) shortage driven by AI infrastructure demand, and analysts don’t expect real relief until 2027 or 2028.
What’s Actually Happening
On August 22, 2026, Bloomberg reported that Nvidia has notified some of its largest customers that prices for servers containing its AI chips — including the Vera Rubin and Grace Blackwell platforms — will rise by more than 15% in many cases. The increases will take effect on systems shipped starting early next year, and the exact amount will depend on the chip generation and memory configuration. Companies that build servers under contract for major data-center operators like Microsoft, Google, and Oracle have reportedly already been informed of the coming increases.
That news landed on top of a consumer graphics card market that was already deep into a price surge that’s been building all year. Nvidia hasn’t officially changed its MSRPs — the company’s website still lists the RTX 5090 at $1,999 and the RTX 5080 at $999 — but actual retail prices have drifted so far from those numbers that the official MSRP has become almost meaningless as a guide to what you’ll actually pay.
Current RTX 50 Series Prices: MSRP vs Reality
Here’s how far street prices have moved from Nvidia’s official launch pricing, based on median Newegg listings tracked in August 2026:
| Model | Launch MSRP | Median Street Price (Aug 2026) | Increase |
|---|---|---|---|
| RTX 5090 | $1,999 | $4,700 | +135% |
| RTX 5080 | $999 | $1,500 | +50% |
| RTX 5070 Ti | $749 | $1,100 | +47% |
| RTX 5070 | $549 | $900 | +64% |
| RTX 5060 Ti 16GB | $429 | $805 | +88% |
| RTX 5060 | $299 | $470 | +57% |
Premium partner cards push even further past these medians. Best Buy has listed the ASUS ROG Astral RTX 5090 as high as $4,829.99, while an ASUS ROG Astral RTX 5080 OC recently sold for $2,099.99 — more than the RTX 5090’s original $1,999 launch price, for a card that launched at $1,499.99 a year earlier. It’s the first time a standard RTX 5080 model has sold for more than the RTX 5090’s MSRP at a major US retailer.
Why Are GPU Prices Rising?
The short version: memory. Every modern GPU relies on high-speed VRAM — GDDR7 on gaming cards, HBM on AI accelerators — and the same manufacturers that supply that memory (Samsung, SK Hynix, and Micron) are also the ones supplying the memory chips going into AI data centers. Cloud providers and AI companies have been buying up memory production capacity at a pace that’s outstripping supply, and manufacturers have no incentive to lower prices when AI infrastructure demand is absorbing everything they can produce.
A single GDDR7 module reportedly costs around $70 on the open market as of August 2026. For a card like the RTX 5080, which uses 16GB of VRAM, that works out to well over $370 in memory cost alone — a sum that, a year ago, would have covered most of the card’s entire bill of materials. Higher-end cards with more VRAM are hit hardest, which is exactly the pattern showing up in the pricing data: the RTX 5090 with its 32GB of GDDR7 has seen the steepest percentage increases of any card in the lineup.
Samsung has separately raised its advanced foundry prices by up to 15% to reflect surging AI demand on its 4nm production lines, according to industry reporting — a move that ripples through to any chip, including GPUs, manufactured using that capacity.
The New Nvidia Data-Center Price Hike, Explained
The August 22 Bloomberg report is specifically about Nvidia’s data-center business, not its consumer graphics cards directly — though the two are connected by the same underlying memory shortage. Nvidia’s Vera Rubin and Grace Blackwell platforms are the chips that power AI training and inference at scale for companies like Microsoft, Google, Amazon, and Oracle. A 15%-plus price increase on those systems represents a meaningful new cost for hyperscalers already spending tens of billions of dollars on AI infrastructure buildouts.
For everyday PC gamers, this specific increase won’t show up as a direct line item — you’re not buying Grace Blackwell servers. But it matters for two reasons. First, it confirms the memory shortage driving consumer GPU prices isn’t easing; if anything, Nvidia’s own supplier costs are still climbing enough to justify a fresh round of hikes on its most profitable product line. Second, when data-center demand for memory and fab capacity increases, that capacity isn’t available to prioritize consumer-facing products like gaming GPUs, which can indirectly worsen availability and pricing further down the stack.
Is AMD Raising Prices Too?
Yes, though generally by less than Nvidia. AMD has confirmed at least a 10% price increase on its Radeon lineup, and reporting through 2026 has tracked several rounds of increases across the Radeon RX 9000 series. That said, AMD’s cards have held up somewhat better against the memory shortage than Nvidia’s higher-VRAM models — the Radeon RX 9070 and RX 9070 XT, both with 16GB of VRAM, have seen less severe inflation than Nvidia cards in the same VRAM tier, and AMD has periodically been available closer to its own MSRP than Nvidia’s equivalent cards.
Timeline: How We Got Here
The RTX 50 series didn’t start out this expensive. Nvidia announced the lineup at CES 2025 with what looked at the time like reasonable pricing: $1,999 for the RTX 5090, $999 for the RTX 5080, $749 for the RTX 5070 Ti, and $549 for the RTX 5070. Supply was tight from day one — the RTX 5090 and RTX 5080 sold out on their January 30, 2025 launch day — but prices stayed roughly close to MSRP through much of the card’s first year.
The shift accelerated through 2026:
- Early 2026: A first wave of price increases in the 10–15% range hit the lineup as memory costs began climbing.
- Mid-2026: A narrower, more targeted increase focused specifically on the RTX 5090, driven by rising GDDR7 costs on its 32GB configuration.
- Summer 2026: A further 20–30% round of increases pushed prices even further from MSRP, with premium AIB (add-in board partner) models leading the way.
- August 2026: Median Newegg prices jumped as much as 39% in a single month compared to June, with the RTX 5060 Ti 16GB and RTX 5070 hit hardest.
- August 22, 2026: Nvidia’s data-center price hike, reported by Bloomberg, signals the underlying cost pressure isn’t letting up.
Notably, Nvidia’s official MSRPs have not moved at any point in this timeline. Every increase has come through wholesale pricing to board partners and retailers, which is why the gap between “official” and “actual” pricing has grown so wide.
When Will GPU Prices Come Down?
Not soon, according to most industry forecasts. Intel has pointed to 2028 before supply conditions normalize. Silicon Motion has warned that the shortage spanning DRAM, NAND, and HBM will likely persist into 2028. Gartner’s own forecasts put relief on the storage side no earlier than 2027. None of the major memory manufacturers have signaled plans to significantly ramp up production in a way that would meaningfully ease prices in the next several quarters, since AI infrastructure spending shows no sign of slowing down.
There was one small, symbolic exception: Nvidia briefly sold Founders Edition RTX 5090, RTX 5080, and RTX 5070 cards at their original MSRP during QuakeCon 2026 in Texas (August 6–9), through a program it called “Verified Priority Access IRL.” The fact that buying an MSRP-priced GPU in person at a gaming convention became notable enough to make tech news headlines says a lot about how far removed MSRP has become from everyday retail reality.
Should You Buy a GPU Now or Wait?
There’s no single right answer here — it depends on how urgently you need to upgrade and which tier of the market you’re shopping in.
- If your current GPU still works and you can wait: Waiting doesn’t guarantee lower prices given the current trajectory, but it does let you watch for temporary dips, open-box deals, or promotional MSRP sales like the QuakeCon event, which may recur at future gaming conventions or sales periods.
- If you need to buy now: Mid-tier cards like the RTX 5070 Ti or RTX 5080 currently offer a better price-to-performance ratio than the RTX 5090, which has effectively priced itself out of the gaming market and into workstation/AI territory for most buyers.
- If you’re flexible on brand: AMD’s Radeon RX 9070 and RX 9070 XT have generally held closer to their MSRP than comparable Nvidia cards with similar VRAM, making them worth a look if you’re not tied to Nvidia-specific features like DLSS.
- If you’re budget-constrained: A previous-generation card (RTX 40-series or older Radeon RX 7000-series) bought used or discounted may offer better value than a current-generation card at inflated pricing, especially since the RTX 5060 Ti and RTX 5060 have seen some of the steepest percentage increases of the entire lineup this month.
Best Value Picks Given the Current Price Surge
Given how distorted current pricing is relative to launch MSRP, the traditional “best value” tier has shifted. The RTX 5080 currently offers roughly 75–80% of the RTX 5090’s gaming performance for well under a third of its street price, making it the most defensible high-end pick right now. The RTX 5070 Ti has remained comparatively stable and sits in a reasonable middle ground for 1440p and entry-level 4K gaming. On the AMD side, the Radeon RX 9070 XT is worth cross-shopping directly against the RTX 5070 Ti, since AMD’s pricing has generally moved less aggressively even with the same 16GB VRAM tier.
If you’re weighing a full system upgrade rather than just a GPU swap, it’s worth checking current pricing across the rest of your build too — memory shortages haven’t been limited to graphics cards. If you’re shopping for a complete gaming rig rather than an upgrade, our ASUS ROG Zephyrus Duo review covers how these cost pressures are showing up in prebuilt gaming laptop pricing as well.
How This Affects Laptops and Other Electronics
Graphics cards aren’t the only category feeling the squeeze. The same DRAM and memory shortage driving GPU prices has already pushed console prices up — Microsoft raised Xbox prices by $100 to $150 earlier this year, citing the same underlying memory cost pressures, and Sony and Nintendo have made similar moves on their own hardware. Gaming laptops with discrete GPUs are exposed to the same VRAM cost increases as desktop cards, layered on top of general DRAM and storage price increases affecting the rest of the system.
MSI, one of the largest GPU board partners, told investors it’s facing a roughly 20% GPU supply shortfall from Nvidia and warned that 2026 will be its toughest year yet because of DRAM, storage, and GPU supply shortages tied to AI demand. The company said it expects the broader PC market to shrink by as much as 20% as a result, and that it’s shifting production away from cheaper, lower-margin components toward mid-range and high-end hardware where it can better absorb rising costs.
What This Means for PC Gaming Going Forward
The practical effect for gamers is that the traditional GPU upgrade cycle — buy a new flagship card every generation or two at a predictable price point — has effectively broken down for now. Price-to-performance comparisons that used to be straightforward now require checking current street prices rather than relying on official specs sheets, since the gap between MSRP and reality can be enormous depending on the exact model and retailer.
It’s also worth noting this isn’t purely a graphics card story — it’s a downstream effect of the broader AI infrastructure boom reshaping demand for the same underlying components used in consumer electronics. As long as AI companies continue expanding data-center capacity at the current pace, memory manufacturers have little incentive to redirect production toward lower-margin consumer parts, which is the core reason analysts aren’t projecting meaningful relief before 2027 at the earliest.
How to Actually Track a Fair Price Before You Buy
Because official MSRP is now such an unreliable guide, the more useful approach is checking a live price tracker or aggregator before committing to a purchase, rather than anchoring to the number printed on Nvidia’s or AMD’s own product pages. Sites that track median and low prices across multiple retailers (Newegg, Best Buy, Amazon, Micro Center) give a far more accurate sense of what you’ll actually pay than any single listing, since prices can swing by hundreds of dollars between retailers for what is nominally the same card.
A few practical habits worth adopting while this market stays volatile:
- Compare the Founders Edition price to AIB partner cards. Nvidia’s own Founders Edition cards have generally stayed closer to MSRP than third-party AIB models from ASUS, MSI, and Gigabyte, which carry premium coolers and factory overclocks that inflate pricing further.
- Watch for temporary MSRP windows. Nvidia’s QuakeCon promotion showed that brief MSRP-priced drops can happen around gaming events or sales periods, even in an otherwise inflated market.
- Don’t assume the newest card is the best value. With this much distortion in pricing, a well-priced previous-generation card can easily beat a current-generation model on price-to-performance, even accounting for the newer card’s feature advantages.
- Factor in your local market. Price increases haven’t hit every region equally — South Korea has reportedly seen RTX 50-series prices climb as much as 30%, while some European markets have seen steeper increases on specific models than the US.
Why This Round of Price Hikes Feels Different
PC hardware has gone through price spikes before — the cryptocurrency mining boom of 2017–2018 and again in 2021 sent GPU prices soaring as miners bought up stock faster than manufacturers could restock shelves. But this current surge has a different underlying cause, and that distinction matters for how long it’s likely to last.
Mining-driven price spikes were tied to the profitability of a specific, relatively volatile activity: when cryptocurrency prices fell or mining algorithms changed, demand collapsed quickly and GPU prices followed. The current shortage is different because it’s driven by structural, multi-year AI infrastructure investment from some of the world’s largest and best-capitalized companies — Microsoft, Google, Amazon, and Oracle aren’t going to suddenly stop building data centers the way individual crypto miners stopped buying GPUs when profitability dried up. That’s a major reason analysts are forecasting a longer recovery timeline this time around, stretching into 2027 and 2028 rather than resolving within a single year.
What Component Makers Are Saying
Beyond MSI’s warning about a 20% supply shortfall and a potential 20% contraction in the broader PC market, other component makers have echoed similar concerns throughout 2026. Silicon Motion, which makes storage controllers, has pointed to a shortage spanning DRAM, NAND, and HBM that it expects to persist into 2028. Gartner’s market research has projected that the storage-specific portion of the crunch will run into 2027 at minimum. None of these forecasts suggest a quick fix, and all of them point to the same root cause: memory manufacturers are prioritizing the AI data-center customers willing to pay premium prices for guaranteed supply, leaving consumer electronics markets to compete for whatever capacity remains.
Frequently Asked Questions
Why did Nvidia raise GPU prices in August 2026?
Nvidia notified major data-center customers on August 22, 2026 that AI server prices will rise more than 15% due to soaring memory chip costs. This follows a broader consumer GPU price surge throughout 2026 driven by the same underlying DRAM and GDDR7 memory shortage.
How much has the RTX 5090 price increased?
The RTX 5090 launched at $1,999 but had a median street price of around $4,700 as of August 2026 — roughly 135% above its original MSRP. Some premium partner models have listed for close to $5,000.
Has Nvidia officially raised its MSRP?
No. Nvidia’s website still lists the original launch MSRPs for the entire RTX 50 series. The price increases have come through higher wholesale pricing to add-in board partners and retailers, not through official MSRP changes.
Is AMD also raising GPU prices?
Yes, AMD has confirmed at least a 10% increase on its Radeon lineup, though its cards have generally seen less severe price inflation than comparable Nvidia models in the same VRAM tier.
When will GPU prices come back down?
Most industry analysts don’t expect meaningful relief before 2027, with some forecasts, including from Intel, pointing to 2028 before supply and pricing conditions fully normalize.
Should I buy a graphics card right now?
If you need one urgently, mid-tier cards like the RTX 5080 or RTX 5070 Ti currently offer better value than the RTX 5090. If you can wait, prices aren’t guaranteed to fall soon, but watching for promotional MSRP sales or open-box deals can help offset the current market conditions.
Why are memory chip prices driving GPU prices up?
GPUs rely on the same DRAM and high-speed memory that AI data centers need for training and inference. As AI companies buy up memory production capacity, less is available for consumer electronics, and manufacturers have little incentive to lower prices while AI demand absorbs their entire output.
Are gaming laptops affected by the same price increases?
Yes. Gaming laptops with discrete GPUs face the same VRAM cost pressures as desktop cards, on top of general DRAM and storage price increases affecting the rest of the laptop’s components.
