Yes, NVIDIA is reportedly raising the cost of GPU-and-memory kits sold to board partners, and higher retail prices are appearing for some GeForce cards. But that is not the same as a confirmed, across-the-board increase to NVIDIA’s official U.S. suggested prices. The distinction matters when deciding whether a card is overpriced, whether to wait, or whether another GPU makes more financial sense. Price observations below reflect U.S. reporting available as of August 16, 2026; listings and availability can change quickly.
What NVIDIA is reportedly raising—and what it has not clearly raised
Supply-chain reporting says NVIDIA has increased prices for GPU-and-memory kits supplied to add-in-board (AIB) partners—the companies that make GeForce cards under brands such as ASUS, MSI, Gigabyte, PNY, and Zotac. Reports describe increases affecting GDDR7 products and potentially some GDDR6 products as well. This is partner-level pricing, not proof that NVIDIA has publicly reset every GeForce card’s manufacturer’s suggested retail price (MSRP). PC Gamer’s reporting on partner kit costs and SemiMedia’s account of the reported change describe the distinction; a universal public MSRP increase is not established by those reports.
Three prices that are easy to confuse
- Launch MSRP: NVIDIA’s suggested reference price when a product launches. It is a benchmark, not a promise that a retailer has stock at that price.
- Partner kit cost: The cost of components supplied to a card maker. A higher kit cost can raise the partner’s cost before the finished card reaches a store.
- Street price: The amount a buyer can actually pay. It reflects the specific card design, partner and retailer margins, shipping, taxes, stock, and seller demand.
The path can be simple—NVIDIA kit cost → board-partner cost → retailer price → consumer price—but each step can add or absorb some of the increase. That is why a higher price in a store does not, by itself, show that NVIDIA changed its published MSRP.
Published RTX 50-series launch prices
NVIDIA’s launch announcements provide the reference prices below. These are U.S. launch MSRPs or starting prices, not guarantees of current availability or current retail prices. NVIDIA’s RTX 50-series product page does not consistently show a current price for every card.
Free tools Windows power users keep installed
One-click scans. No signup required.
#1 Best Overall
- AI Performance: 767 AI TOPS
- OC mode: 2632 MHz (OC mode)/ 2602 MHz (Default mode)
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Axial-tech fan design features a smaller fan hub that facilitates longer blades and a barrier ring that increases downward air pressure
- A 2.5-slot design maximizes compatibility and cooling efficiency for superior performance in small chassis
| GPU | Published launch price | Source |
|---|---|---|
| GeForce RTX 5090 | $1,999 | NVIDIA RTX 50-series announcement |
| GeForce RTX 5080 | $999 | NVIDIA RTX 50-series announcement |
| GeForce RTX 5070 Ti | $749 | NVIDIA RTX 50-series announcement |
| GeForce RTX 5070 | $549 | NVIDIA RTX 50-series announcement |
| GeForce RTX 5060 Ti 16GB | $429 | NVIDIA RTX 5060 family announcement |
| GeForce RTX 5060 Ti 8GB | $379 | NVIDIA RTX 5060 family announcement |
| GeForce RTX 5060 | $299 | NVIDIA RTX 5060 family announcement |
| GeForce RTX 5050 | $249 starting price | NVIDIA RTX 5050 announcement |
Which cards appear most exposed to higher prices?
Reports point to GDDR7-equipped RTX 50-series cards as a central part of the partner-cost story. NVIDIA lists the RTX 5090, 5080, 5070 Ti, 5070, and 5060 family with GDDR7 memory in its GeForce specifications comparison. A separate report says the alleged kit-price changes may also reach GDDR6 products, including the RTX 5050 and some older GeForce cards; treat that broader scope as reported rather than as a public NVIDIA policy. Tech Times reported on the possible wider product coverage.
Higher-memory and high-demand models
Cards with larger memory configurations, costly components, strong demand, or limited stock have more ways to be affected. Recent price coverage has called out the RTX 5070, RTX 5070 Ti, RTX 5080, and RTX 5060 Ti 16GB among models with notable price pressure. That does not mean every card in those families has the same markup: an entry-level partner model, a premium cooler design, and a marketplace listing can have very different prices.
Older cards, laptops, and workstation products are separate cases
Reports about GDDR6 kit costs do not establish the price of every older GeForce model. Laptop RTX GPUs also are not directly comparable with desktop cards carrying the same model number; power limits and cooling affect performance. And the RTX Pro 6000 Blackwell is a workstation product, not evidence of a GeForce MSRP change. Tom’s Hardware’s RTX Pro 6000 pricing report concerns that separate product category.
Rank #2
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5070 Ti
- Integrated with 16GB GDDR7 256bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
How much have retail prices moved?
There is no single current price for a GPU family. The figures below are reported U.S. retail signals, not new NVIDIA MSRPs. They refer to particular tracking or listings available in August 2026, and should not be read as a nationwide average or a guaranteed price. The exact seller, model, stock status, and whether a listing is new all matter.
| GPU | Published launch reference | Reported U.S. retail signal | What the signal shows |
|---|---|---|---|
| RTX 5070 | $549 | About $700 in recent tracking, compared with roughly $550 previously | PC Gamer reported a rise near $150; this is observed retail pricing, not an MSRP reset. Source |
| RTX 5090 | $1,999 | Some reported listings around $4,381–$4,399 | These are high-priced listings, not the typical price or an official NVIDIA price. Source |
| RTX 5080 | $999 | Some premium models reported around $2,099 or higher | A premium partner-model listing does not represent every RTX 5080. Source |
| RTX 5070 Ti | $749 | Partner models reported materially above launch reference | No single comparable street price is established here; check the exact model and seller. |
| RTX 5060 family | $299–$429, depending on model | Some models reported closer to launch reference, particularly lower-memory versions | Availability and exact card design determine whether a listing is a genuine near-MSRP option. |
Tom’s Hardware reported median U.S. RTX 50-series price increases of up to 39% in Newegg listings; its reporting also described a 36% median increase for the RTX 5070 and a 27% increase for the RTX 5060. These are retailer-listing medians for the coverage period, not prices paid by every buyer or increases in NVIDIA’s official MSRP. See Tom’s Hardware’s listing-based price analysis.
Why consumer prices can rise
Memory costs can feed into a graphics card’s cost
GDDR6 and GDDR7 memory are components of a graphics card. If memory-inclusive kits cost partners more, the finished board can become more expensive to produce. SemiMedia linked the reported kit pricing to climbing GDDR costs. Its report describes that connection.
Rank #3
- 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
AI demand can tighten the broader supply chain
Reporting cites AI-related demand as a contributor to pressure on memory and semiconductor capacity. The careful version of the claim is that demand for AI hardware can intensify competition for manufacturing resources and components; it does not mean every GDDR chip used in a gaming card is being diverted directly into an AI accelerator. Tom’s Hardware’s GPU price tracker discusses AI-driven memory demand in the context of GPU pricing.
Supply, partner pricing, and retailer inventory add further layers
When gaming-card supply is tight and buyers remain interested, retailers have less reason to discount. A partner can also price a premium cooler or factory-overclocked design above a basic model. After lower-cost inventory sells through, a store may be left with more expensive versions or marketplace sellers. These factors can make a reported supplier-cost change show up gradually and unevenly in consumer listings.
Some coverage has also suggested that NVIDIA prioritizes higher-margin products, including AI hardware and certain high-end gaming cards. That is a reported interpretation of supply allocation, not a confirmed explanation from NVIDIA. Tom’s Hardware’s coverage discusses the supply and pricing context.
Rank #4
- Powered by the NVIDIA Blackwell architecture and DLSS 4
- Powered by GeForce RTX 5060
- Integrated with 8GB GDDR7 128bit memory interface
- PCIe 5.0
- WINDFORCE cooling system
Who is responsible for the price a buyer sees?
There is no single actor setting every part of the final price. Higher component costs can be real, but how much of them gets passed on—and how much extra is added—depends on decisions throughout the channel.
- NVIDIA: Reportedly sets higher prices for kits sold to partners; that does not establish a public GeForce MSRP increase.
- Board partners: Set prices for their finished cards, including differences for cooling, power design, and factory overclocking.
- Retailers and marketplace sellers: Set listing prices based on wholesale costs, inventory, competition, and demand.
- Buyers: Face the combined effect, but can compare models, wait, or choose another route rather than treating every asking price as unavoidable.
Cost pressure and pricing power can coexist. Memory may cost more while a particular partner, retailer, or seller also chooses to charge a larger premium. A high listing alone cannot tell you how much of the total comes from each step.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Should you buy an NVIDIA GPU now or wait?
Make the decision around your need, the exact card price, and the value of NVIDIA-specific features—not around a prediction that prices must fall by a certain date.
Best Value
- Powered by the NVIDIA Blackwell architecture and DLSS 4 OC mode: 2640MHz/Default mode: 2610MHz (Boost Clock)
- Military-grade components deliver rock-solid power and longer lifespan for ultimate durability
- Protective PCB coating helps protect against short circuits caused by moisture, dust, or debris
- 3.125-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
Buy now if the card solves a real need at a price you can justify
- You need the GPU for work, rendering, development, or a game your current card cannot handle.
- You find the exact model in stock from a reputable seller at or near its launch reference, or the premium is reasonable for its cooler, warranty, or form factor.
- The cost of waiting—lost work time or an unusable PC—matters more than a possible future discount.
Wait if your current card is adequate and the markup is large
- You are targeting an RTX 5070, 5070 Ti, 5080, or 5090 well above its launch reference.
- You can wait for another restock, a promotion, or a better comparison with competing products.
- You do not rely on CUDA, OptiX, or other NVIDIA-specific application support that makes switching costly.
Compare by workload, not just brand or VRAM headline
For gaming, consider whether a lower-tier card can meet your resolution and settings needs. NVIDIA promotes DLSS 4 and Multi Frame Generation for RTX 50-series cards, but generated frames are not the same as higher native rendering performance; decide whether those features matter for the games you play. NVIDIA’s RTX 5060 family announcement discusses its DLSS features.
For raster-focused gaming, compare current AMD and Intel options if you do not need CUDA or OptiX. AMD’s Radeon product page and Intel’s Arc graphics page identify their product lines, but prices and value depend on live availability and the exact competing model. For creator software, local AI, or professional applications, check compatibility before choosing a lower-priced alternative.
Do not choose an 8GB card solely because it is the cheapest option if your games, texture settings, or applications need more memory. The reverse is also true: paying a large premium for 16GB or 32GB solely as future-proofing may not make sense if your actual workload does not use it.
Used, previous-generation, and prebuilt options
A used or older-generation card can be worthwhile when the price is substantially below a new equivalent and its condition, warranty, and performance fit your needs. Check the seller’s return terms and, where possible, test the card. Scarcity can inflate older-card prices too, so a previous generation is not automatically a bargain.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
A prebuilt gaming PC is another comparison when a standalone GPU is heavily marked up. Compare the total price and the CPU, memory, storage, power supply, cooling, warranty, and upgradeability; a low headline price can conceal weak components or a restrictive design. Vendor ranges include Dell desktop PCs, HP OMEN desktops, and Lenovo Legion desktops.
How to avoid overpaying for a graphics card
- Compare like with like. Check the exact GPU, VRAM capacity, partner model, and whether the card is new, refurbished, or used against the launch reference above.
- Check the seller and stock. An MSRP listing may be out of stock, sold by a third party, delayed, limited to store pickup, or subject to a membership requirement. Confirm the seller, delivery date, warranty, and return policy.
- Calculate the checkout total. Add shipping and applicable taxes before comparing prices; regional taxes, import costs, and distribution can make U.S. launch prices poor benchmarks elsewhere.
- Judge premium features honestly. A factory overclock or premium cooler can raise a card’s price, but a small clock-speed difference alone may not justify a large markup. Consider whether the design’s cooling, noise, dimensions, power delivery, or support solves a specific need.
- Set a maximum premium before shopping. Decide what the card is worth for your workload and do not let a dramatic high listing redefine a sensible price.
- Use more than one price observation. A single expensive marketplace listing is not the market price, and one brief MSRP restock does not prove supply has normalized. Check price history and multiple legitimate sellers.
The evidence and price signals here are U.S.-focused. In Canada, the U.K., Europe, and other markets, currency conversion, VAT or sales tax, duties, distribution, and local stock can change the comparison. Do not apply the U.S. dollar figures as local price targets.
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.
