According to a 10-country graphics card price survey that TechSpot published on November 2, GPU prices rose 11% in October from the previous month. The survey tracks the lowest in-stock price for each model in each country. In the US, the RTX 5090 reached $9,400, about 4.7 times its launch manufacturer's suggested retail price (MSRP) of $1,999. Rising memory prices are part of the story, but they don't explain everything. Even among three GeForce cards that all carry 16GB of GDDR7, the lowest US price exceeds MSRP by anywhere from $361 to $1,201. Using TechSpot's data, this article examines how much non-memory factors may be affecting prices.
GPU prices up 11% in October; RTX 5090 hits $9,400 in the US
TechSpot looked at ten countries, including the US, the UK and India. It checked the lowest price for each model on price comparison sites and confirmed that stock was actually available. The benchmark is the cheapest product available for each model, not the average selling price across the market. TechSpot converts the US-dollar MSRP into local currency using exchange rates and tax rates, then calculates how far the lowest shelf price exceeds it.
Using this method, October GPU prices rose 11% from September and 35% over the three months since August. Across all models and all ten countries, the lowest shelf price averaged 44% above MSRP. The increase varies widely by model, however. In the US, the RTX 5060 Ti 8GB was 40% above MSRP, the RTX 5080 was 120% above and the RTX 5090 was 370% above. Meanwhile, the RTX 5060 Ti 16GB was 84% above and the RTX 5050 was 69% above, so higher-performing cards don't simply see larger percentage increases.
The RTX 5090 is the most extreme case. Its lowest US price climbed from $3,500 in February to $6,890 in September and $9,400 in October. The October month-over-month increase was 36% in the US and 32% on the 10-country average, which put the average price 262% above MSRP. Already in February the card was selling at 1.75 times MSRP, so the surge has been under way for some time.
There were reports that the RTX 5090 was being discontinued, but Jeff Fisher, NVIDIA's senior vice president responsible for consumer computing, said in an interview after the "RTX Spark Launch Party" in Taipei that production and shipments continue. According to a report by Taiwan's 4Gamers on October 16, Fisher said the RTX 5090 chip has other variants, while also acknowledging that supply chain capacity is "indeed somewhat constrained." Such supply constraints may also be affecting prices, but the extent is unclear.
Same 16GB of GDDR7, but a $361 to $1,201 gap above MSRP
It is true that upward pressure on memory prices is intensifying. On September 30, research firm TrendForce said supply of GDDR6 is constrained as manufacturers cut production and withdraw from supply, and that GDDR7 is also constrained because advanced-process capacity is being redirected to server DRAM and other products. It forecasts that graphics DRAM prices will keep rising.
So how much does higher memory pricing feed into graphics card prices? Memory cost depends mainly on type and capacity, so cards with the same 16GB of GDDR7 can be assumed to see similar increases in memory cost even if their GPUs differ in performance. This ignores differences in memory speed grade. Instead of percentage increases, we therefore compare the dollar amount by which the lowest US price in October 2026 exceeds the launch MSRP.
Calculated from TechSpot's price table and the MSRPs announced by NVIDIA, the gaps for the three 16GB GDDR7 GeForce cards are as follows.
| Model | Lowest US price, October | Launch MSRP | Amount above MSRP | Above MSRP |
|---|---|---|---|---|
| RTX 5060 Ti 16GB | $790 | $429 | $361 | +84% |
| RTX 5070 Ti | $1,250 | $749 | $501 | +67% |
| RTX 5080 | $2,200 | $999 | $1,201 | +120% |
Even with the same 16GB of GDDR7, the amount above MSRP is larger for higher-end models. The RTX 5070 Ti's premium is $140 larger than the RTX 5060 Ti 16GB's, and the RTX 5080's is $840 larger. If the rise in memory cost is roughly the same across the three cards, a considerable part of this gap likely comes from factors other than memory. However, the costs of the GPU die, board, power circuitry and cooling also differ, so the entire difference can't be attributed to non-memory markups.
The 10-country averages show a similar pattern. In TechSpot's conversion of each country's prices into US dollars, the premium over MSRP was about $300 for the RTX 5060 Ti 16GB, about $450 for the RTX 5070 Ti and about $770 for the RTX 5080. The RTX 5060 Ti 16GB and RTX 5070 Ti carry the same speed of 16GB GDDR7, yet differ by about $150. TechSpot points this out and concludes that memory cost increases alone cannot account for the entire rise.
The RTX 5080's price swing is especially large. Its lowest US price rose 39%, from $1,580 in September to $2,200 in October, and the 10-country average rose 22% month over month. At the time of the survey, however, Newegg in the US had no direct stock, and only third-party sellers at inflated prices remained. TechSpot sees this as one reason for the sharp US increase, so the $2,200 figure may reflect marketplace-specific high prices.
On the RTX 5080's price rise, TechSpot states: "This is not due to memory price increases, plain and simple." It says the cost of 16GB of GDDR7 has not itself risen by $700, and names price gouging, demand from AI uses and the scarcity of the RTX 5090 as possible factors. The size of each factor's effect is unknown.
Citing people in the retail industry, the publication also reports that the RTX 5080's wholesale price has risen sharply in recent weeks and that shelf prices are expected to climb further. It also mentions a rumor that a 24GB version replacing the 16GB model will appear in early 2027, but that is unconfirmed.
The gap between 8GB and 16GB versions has widened more for GeForce
Comparing products that use the same GPU and differ only in memory capacity reveals another trend. The RTX 5060 Ti's 8GB and 16GB versions differed by $50 at launch, and so did AMD's RX 9060 XT. But in October's lowest US prices, that gap widened to $260 for the RTX 5060 Ti and $120 for the RX 9060 XT. AMD's launch prices were announced as suggested retail prices (SEP), which TechSpot treats as MSRP.
Calculating the amount above MSRP from TechSpot's survey results and each product's launch price gives the following.
| Model | Memory | Lowest US price, October | Launch MSRP | Amount above MSRP |
|---|---|---|---|---|
| RTX 5060 Ti 8GB | 8GB GDDR7 | $530 | $379 | $151 |
| RTX 5060 Ti 16GB | 16GB GDDR7 | $790 | $429 | $361 |
| RX 9060 XT 8GB | 8GB GDDR6 | $500 | $299 | $201 |
| RX 9060 XT 16GB | 16GB GDDR6 | $620 | $349 | $271 |
The difference in amount above MSRP between 16GB and 8GB versions is $210 for the RTX 5060 Ti and $70 for the RX 9060 XT. On the 10-country average, the RTX 5060 Ti 16GB was 69% above MSRP, about $300 more, while the 8GB version was 24% above, less than $100. The RX 9060 XT 16GB was 58% above, about $200, and the 8GB version 46% above, about $140. The extra markup for an additional 8GB of capacity is larger for GeForce than for Radeon.
However, the RTX 5060 Ti uses GDDR7 and the RX 9060 XT uses GDDR6, so the memory standards differ. TrendForce's published materials don't give per-unit prices for either standard, so it isn't possible to tell how much of the $210 and $70 difference stems from memory cost. The lowest US price may also shift as the products surveyed change.
Comparing 16GB products by manufacturer also shows differences. According to TechSpot, the RX 9060 XT 16GB is about $200 above MSRP in Radeon's lineup, and the higher-end RX 9070 XT is at a similar level, $216. By contrast, GeForce's RTX 5060 Ti 16GB is over $300 above, and the premium grows with higher-end models. Even the 12GB RTX 5070 showed a markup of more than $260. Because Radeon and GeForce use different memory standards, a simple comparison isn't possible.
Month-over-month increases in October tell a different story. On the 10-country average, the RX 9060 XT 16GB rose 17%, more than the RTX 5060 Ti 16GB's 7%. The RX 9070 XT's 12% also exceeds the RTX 5070 Ti's 8%, though the RTX 5080 rose 22%. TechSpot summarizes that, excluding the RTX 5090, Radeon prices rose more than GeForce prices. If that continues, the gap in cumulative increases since launch between the two brands could narrow.
In Japan, the lowest listed prices are 1.78 to 3.46 times the yen-converted US MSRP
We also checked domestic Japanese prices. Searching model names in Kakaku.com's video card listings and sorting by lowest price, the lowest tax-inclusive listings as of November 3, 2026 were ¥1,198,000 for the RTX 5090, ¥331,800 for the RTX 5080 and ¥230,800 for the RTX 5070 Ti. These are the cheapest listings among those with the model name in the product name, and we did not confirm whether they are in stock.
As the comparison baseline, we used the exchange rate of ¥157.47 to the dollar, the New York noon rate on October 30 published by the Federal Reserve Board (FRB). Converting the US-dollar MSRPs to yen and adding 10% consumption tax gives about ¥346,300 for the RTX 5090, about ¥173,000 for the RTX 5080 and about ¥129,700 for the RTX 5070 Ti. This is a baseline set by our editorial staff for comparison, not an official price announced by NVIDIA for the Japanese market.
Comparing this baseline with Kakaku.com's lowest listings, the RTX 5090 is about 3.46 times, the RTX 5080 about 1.92 times and the RTX 5070 Ti about 1.78 times, equivalent to +246%, +92% and +78% above MSRP. Compared with the US increases (+370%, +120%, +67%), the RTX 5090 and RTX 5080 are lower in Japan, while the RTX 5070 Ti is higher. Against TechSpot's 10-country averages (+262%, +77%, +60%), the RTX 5090 is slightly lower while the other two are higher, so the Japanese market can't be called uniformly expensive. For reference, the US RTX 5090 lowest price of $9,400 comes to about ¥1.48 million at the same exchange rate (excluding sales tax).
These comparisons carry caveats. Kakaku.com's figures are listed prices with stock unconfirmed, a different condition from TechSpot's stock-verified lowest prices. Listings whose product names lack the model name may be missed in search results, so these may not be the lowest prices in the entire domestic market. The retrieval date of November 3 also differs from the timing of TechSpot's October survey.
DRAM shortage expected to widen in 2027, and NVIDIA's earnings don't reveal GeForce's reality
Upward pressure on memory prices is expected to continue. On September 30, TrendForce projected that general-purpose DRAM contract prices in the fourth quarter of 2026 will rise 10% to 15% from the previous quarter. In its July 30 announcement, it estimated the 2026 DRAM supply-demand sufficiency ratio at around −1% to −2%, and forecast that in 2027 demand growth will outpace supply capacity expansion and tighten the market further. However, the published materials give no per-unit prices from which a per-card memory cost could be calculated, so the portion of the recent price rise explained by memory costs can't be derived.
NVIDIA's earnings materials also make it hard to gauge GeForce price trends on their own. Starting with the first quarter of fiscal 2027, announced on May 20, the company changed its market revenue disclosure to two categories: "Data Center" and "Edge Computing." Edge Computing covers PCs, game consoles and workstations as well as AI-RAN base stations, robotics and automotive. First-quarter revenue was $6.4 billion. The CFO commentary said growth was driven by rising demand for Blackwell-based workstation products, while consumer PC demand slowed on higher memory and system prices. In the second-quarter results announced on August 26 (quarter ended July 26), total revenue was $96.2 billion, and Edge Computing was $7.2 billion, up 27% year over year.
The closest information to GeForce results is the disclosure for the Graphics segment, an accounting reporting segment. According to the Form 10-Q for the second quarter of fiscal 2027, revenue in the segment, which includes GeForce and enterprise workstation GPUs, was $7.922 billion, up 46% year over year. Operating income was $3.899 billion, up 74% from $2.242 billion a year earlier. NVIDIA said sales of Blackwell products drove the increase, but it doesn't disclose GeForce revenue, unit volume or average selling price on their own. It is therefore unclear whether the growth came from higher prices or higher volume. The materials we reviewed were the 10-Q, the earnings release and the CFO commentary; we did not verify remarks made on the earnings call.
AMD's situation is similar. Its Gaming segment revenue in the second quarter of 2026 (ended June 27) was $779 million, down 31% year over year. But the main cause was lower sales of semi-custom products for game consoles, so Radeon supply and demand can't be judged from that figure. On the supply side too, NVIDIA's Fisher acknowledged capacity is "somewhat constrained" but gave no specific figures.
If the memory supply-demand squeeze lasts into 2027, it is hard to expect costs to fall quickly. That makes the price gaps between models carrying the same memory a useful clue for reading the market. If the $140 gap between the RTX 5070 Ti and RTX 5060 Ti 16GB, or the $840 gap with the RTX 5080, narrows going forward, non-memory factors in high-end pricing may have weakened. If the gaps keep widening, high-end GeForce price movements can't be explained by memory market supply and demand alone. TechSpot's ongoing 10-country survey will be an important reference for tracking real selling-price changes that NVIDIA's earnings disclosures don't reveal.