www.techspot.com
It's time to cut through the BS and talk about graphics card value based on real-world prices. You've seen the reviews, you know how the latest GPUs stack up based on MSRP, and you also know how those MSRPs are basically fake in many instances.In most regions around the world, you just can't walk into a store and buy a card for the price that AMD and Nvidia advertised, making those prices irrelevant at least for now, while the market is severely disrupted. With this in mind, we need to talk about which graphics cards are the best and how they compare based on realistic, real-world prices we're seeing in the market today.And that's exactly what we're doing in this article. We'll be looking at real pricing in multiple regions to properly compare the GeForce RTX 50 series and Radeon RX 9000 series to each other, as well as to prices you could have paid for graphics cards not too long ago.Sure, these companies can advertise prices to get better reviews and then not deliver on those promises, assuming everyone will simply forget about a blatant bait-and-switch. Well, not us, because we continually evaluate the market, and it's safe to say we're not happy with what we've found.Most Models are Out of StockAnalyzing pricing right now is trickier than usual because the vast majority of models simply aren't available at all. So if we did a review based on current prices for in-stock models, it would be pretty boring, featuring just a handful of comparisons. But it's not like graphics cards aren't being sold at all retailers are selling models almost as soon as they arrive.The major complaint we've seen from our readers is that these restocks virtually never include graphics cards priced at MSRP. They're always overpriced mid-range or premium models, sometimes priced even higher than during the last restock. Anything remotely close to a reasonable price sells out instantly, and even cards we'd consider "unreasonable" disappear quickly too.Our Data PointsTo factor this problem into our value discussion, we're looking at graphics card offerings in a unique way. We've examined the line-ups from each company and their listed prices at major retailers.We've discarded the lowest-tier model entirely from price evaluation, as in many cases these cards either don't exist or aren't restocked in meaningful quantities. Then we've selected some data points from the rest of the line-up.The next model up from the base version becomes our baseline real-world price for that graphics card. In a normal market, these slightly upgraded models are listed with a small premium over MSRP, usually around 5%. But the market is far from normal right now, so the data will be quite revealing.The second data point we're looking at is the price for a premium GPU the most expensive card in the lineup, excluding liquid-cooled models. Normally, these cards would cost around 10 15% above MSRP. But in a heavily supply-constrained market, this data point is important because it represents the price you might have to pay if GPU makers decide to supply only premium models rather than lower-priced cards.After all, if an AIB can only produce 100 RTX 5090s and knows they'll all sell instantly, why would they produce anything other than the most expensive GPUs? And we're certainly seeing restocks heavily biased toward these premium models.For example, this is the RTX 5070 Ti lineup from MSI in Australia. We're discarding the price of the Shadow 3X OC model as it's an unrealistic MSRP card that's not actually available. The next model up is the Ventus 3X OC, which becomes our baseline real-world price.Then we're also noting the price of the Vanguard SOC, the premium model. We repeat this process for all other AIBs and average the prices for these slightly upgraded and premium cards, and these averages go into our cost-per-frame graphs.For older GPUs in the GeForce RTX 40 and Radeon 7000 series, we're using real-world pricing for these cards prior to stock drying up at the end of 2024. As we explained in our recent GPU pricing feature, previous-gen models were discontinued earlier than expected due to anticipated GPU launches, which were ultimately delayed. This cleared out the market and caused prices to rise for remaining stock.Therefore, the few remaining cards today don't fairly represent what those cards cost throughout their lifespan. We want to compare the real price of new models to the real price of previous models at a time when they were readily available. This way, Nvidia and AMD can't simply raise prices across the board right before launching disgustingly overpriced new models.What Normal MSRP GPU Value Would Look LikeAs a quick refresher, here's what the line-ups should look like based on MSRP. The Radeon RX 9070 XT should offer a 19% lower cost per frame than the GeForce RTX 5070 Ti at 4K. The RX 9070 should be 8% cheaper per frame than the RTX 5070. The RTX 5080 should be 6% cheaper per frame than the RTX 4080 Super, and so on...These margins shift slightly at 1440p compared to 4K. For example, at 1440p, the RX 9070 XT reduces to just a 15% lower cost per frame compared to the RTX 5070 Ti. Nonetheless, this still places the Radeon model in a solid position.Cost Per Frame: Real U.S. PricingNow let's take a look at real-world pricing for the next model up in each lineup, starting with the 4K data in the current U.S. market. For the latest generation of cards, this data is based on listed prices at Newegg. As we can see, many models have drastically increased in price.The RTX 5080 becomes a $1,350 GPU, the RTX 5070 Ti increases to $900, and the RTX 5070 jumps to $660. AMD's cards aren't immune either: the 9070 XT is essentially a $750 card, and the 9070 is about a $670 card under the pricing evaluation system we described earlier.So how does this affect value? Well, directly comparing current-generation models, it basically maintains the status quo. At 4K, the RX 9070 XT remains 16% cheaper per frame than the RTX 5070 Ti using realistic pricing, as both models are currently around $150 above their MSRP. We see a similar trend with the 9070 versus the 5070, as both cards are roughly the same price right now.Where the comparison becomes concerning is when comparing current-generation models to prices paid for previous-generation GPUs throughout most of last year. The RX 9070 XT, for example, ends up offering nearly the same value as the 7900 XT, which was regularly available for around $700.For Nvidia GPUs, it's even worse the RTX 5070 Ti shifts from being slightly better value than the RTX 4070 Ti Super to becoming 5% more expensive per frame. The RTX 5080 is a massive 30% more expensive per frame than the RTX 4080 Super, positioning it as even worse value than the RTX 4080 and making it a terrible purchase in the current market.And good luck if you're gaming at 1440p. Here, the RTX 5080 is 35% more expensive per frame compared to the typical price of the RTX 4080 Super in 2024. The RX 9070 XT also offers poor value, as it's unable to deliver a true next-generation experience due to being 22% more expensive per frame than the RX 7800 XT, and 5% more expensive per frame than the 7900 XT.At MSRP, the 9070 XT would offer better value than both previous-gen models, but at current real-world pricing, there's only a slim 12% discount per frame compared to the RTX 5070 Ti.Things get even worse when looking at the worst-case scenario for GPU pricing, particularly when comparing current models to older models. Similar margins remain between AMD and Nvidia: under premium pricing, the 9070 XT is 17% cheaper per frame than the 5070 Ti, compared to 16% using the next-model-up pricing and 19% at MSRP.But in general, you definitely don't want to pay these prices, as a 9070 XT becomes 4% more expensive per frame than a 7900 XT, the 5070 Ti is 13% more expensive than the 4070 Ti Super, and the 5080 is a staggering 41% more expensive per frame than the 4080 Super presenting even worse value than the RTX 4090.Consumers are being severely overcharged if the only available stock is premium models; these GPUs simply shouldn't be purchased or even exist at the current list prices we're seeing today.Cost Per Frame: Real Australian PricingOf course, the United States is just one side of the story. The dynamics at play can be totally different in other regions, so let's look at Australia. We examined real list pricing for each lineup at PC Case Gear and Scorptec to create these average prices.Essentially, new GPUs are a terrible deal in Australia, especially if you're interested in high-end Nvidia GPUs. Even the next tier up from the base RTX 5080 forces you to spend around $2,550 AUD, making it worse value than the RTX 4090 at its typical price in mid-2024.The company benefiting the most here is AMD. The RX 9070 XT, in a direct comparison with the RTX 5070 Ti, offers a 24% lower cost per frame, so at least in Australia, the Radeon card presents better value than in other regions at realistic market prices. This likely explains why RTX 5070 Ti models have started reappearing in stock and sitting on shelves priced at $1,700 or more because even $1,700 isn't particularly attractive.What's interesting is that, in the battle between the RTX 5070 and RX 9070, we see much better availability compared to other cards. Therefore, we've listed the cheapest available prices here. The RTX 5070 is available at its MSRP of $1,110 AUD, which is a generally poor deal, while the RX 9070 at a similar price actually ends up providing better value than the typical RX 9070 XT pricing. Currently, it's 7% cheaper per frame, although this margin will likely shrink if the RX 9070 XT returns to stock closer to MSRP since the RX 9070 has little room for price movement.While these comparisons between current models are informative, overall pricing is disappointing compared to previous-generation GPUs. Here in Australia, the RX 9070 XT is 13% more expensive per frame than the RX 7900 XT, which was regularly priced between $1,100 and $1,200 AUD as recently as January. Realistically, the RX 9070 XT should be priced below $1,000 to make sense.Even the better-priced RX 9070 still offers worse value than the RX 7900 XT. Remember, we're strictly discussing rasterization performance here, not other factors like ray tracing and FSR 4.On Nvidia's side, the situation is even worse. The realistic price of the RTX 5070 Ti makes it 27% more expensive per frame compared to the RTX 4070 Ti Super, which was commonly available for around $1,300 AUD in 2024. Based on MSRP comparisons, the RTX 5070 Ti should actually be cheaper than the 4070 Ti Super, even considering currency exchange rate fluctuations. In Australia, this GPU realistically should be around $1,300, not nearly $1,800. And let's not even discuss the RTX 5090.When considering premium models, pricing gets even more absurd, especially for high-end GeForce cards that make absolutely no sense. Who would pay nearly $6,000 AUD for an RTX 5090? That's outrageous. The RX 9070 XT ends up 26% cheaper per frame than the RTX 5070 Ti, so Radeon GPUs still maintain better relative value locally even if only premium models see resupplies. Of course, "better" is relative, because compared to the RX 7000 series, current buyers are still getting significantly worse deals.Cost Per Frame: Real European PricingFinally, let's take a look at the European market or at least part of it. This data comes from German retailers. While we could have included many other countries, the German market provides a clear picture of broader trends.Unfortunately, things look rough for our European friends, particularly for those interested in GeForce 50 series GPUs, which remain excessively expensive. The price for the next model up in the RTX 5070 Ti lineup is 1,040, making it 11% more expensive per frame than the RTX 4070 Ti Super, which was regularly available around 860 in mid-to-late 2024. Essentially, German consumers are paying more than Americans for the RTX 5070 Ti but aren't being hit as hard as Australians.We see similar pricing trends for the RTX 5080 and RTX 5090, both extremely overpriced. In each case, these cards offer much worse value compared to the RTX 4080 Super and RTX 4090, respectively. The RX 9070 XT is also priced poorly, with a 12% higher cost per frame compared to the RX 7900 XT's typical price from just a few months ago quite similar to Australia's situation.As for the RTX 5070 and RX 9070, we're seeing similar trends in Germany as in Australia, with these GPUs available at prices between their supposed MSRP and the next model up. Both cards, which have a $550 MSRP, should ideally cost around 600 in Germany but are currently priced approximately 100 above that. As a result, they're poor value compared to previous-generation models. For instance, the RTX 5070 is 10% more expensive per frame compared to the RTX 4070 Super, but at least these cards are actually purchasable right now.The question then becomes: which card currently offers the best value?Right now, the RX 9070 slightly edges out the listed price for the next-tier RX 9070 XT, delivering a 6% lower cost per frame. We wouldn't call this a great deal, considering it still represents worse value than previous-generation GPUs. However, it's currently 5% cheaper per frame than the RTX 5070 and 22% cheaper per frame than the RTX 5070 Ti. Meanwhile, if the RX 9070 XT were available at its listed price, it would offer 17% cheaper cost per frame compared to the RTX 5070 Ti.If AIBs only supply premium models, the market outlook becomes even worse, with new GPUs providing significantly poorer value compared to last-generation models. For example, the RTX 5080 would become 44% more expensive per frame than the RTX 4080 Super, and currently, if you want to buy an RTX 5080 that's actually in stock, you'll likely pay a price close to this extreme.Clearly, this is a very bad situation. Under premium pricing, the RX 9070 XT emerges as the best value among the new generation GPUs, offering a 19% lower cost per frame compared to the RTX 5070 Ti but still not representing good overall value compared to previous-gen options.What We LearnedOverall, when assessing the graphics card market based on real-world prices, things look quite dire at the moment. Whether we look at the United States, Australia, or Germany, the typical prices of available models make new-generation graphics cards a worse value than their previous-generation counterparts.Of course, in most situations, you currently can't buy either current or previous-generation models easily. But when comparing prices of relatively basic AIB models today to GPU prices just six months ago, it's clear that buyers are being significantly overcharged when purchasing GPUs as they arrive at retailers.This is largely due to an unusually wide gap between the base-model cards listed at MSRP and the next tier in an AIB's lineup. In previous years, it was typical to pay around 5% extra for a second-tier card for example, a GPU with a $750 MSRP would typically see the next model up priced around $790. Instead, we're currently seeing closer to a 20% premium for these models, and these higher-priced variants are the ones being supplied. And if it's not these models, it's often even more expensive ones.In reality, base-model GPUs are the only models worth buying if you want a reasonable deal compared to the 2024 GPU market. The MSRP is the only price point that genuinely makes sense. Perhaps a 5% premium could be justified for instance, a $790 RTX 5070 Ti would still offer slightly better value than the RTX 4070 Ti Super, around 8% cheaper per frame but at a 20% premium over MSRP, it's simply not worth it. At such prices, buyers are getting badly overcharged and would be better off waiting for the market to stabilize.Interestingly, inflated GPU prices appear to have impacted Nvidia and AMD roughly equally in market segments where they directly compete. Prior to launch, we heard arguments suggesting that AMD's Radeon RX 9070 XT price of $600 was significantly more attractive compared to the RTX 5070 Ti, which was selling for over $900 at the time. Some argued we shouldn't compare these GPUs at their advertised MSRP because Nvidia's advertised price is fake and AMD's is real.However, nearly two weeks after the RX 9070 XT launch, if we compare similar models in their respective lineups excluding nonexistent base-model MSRPs we find the price-to-performance differences remain similar. At MSRP, the RX 9070 XT is 19% cheaper per frame than the RTX 5070 Ti. But based on real-world prices today, the RX 9070 XT is only 16% cheaper per frame in the US market, as both models have experienced comparable price hikes.In Germany, the RX 9070 XT ended up 17% cheaper per frame, and Australia benefited more significantly with a 24% discount per frame. If the RX 9070 XT had actually remained at $600 as some anticipated while the RTX 5070 Ti stayed around $900, the RX 9070 XT would have offered a 32% cost-per-frame advantage. Unfortunately, this scenario hasn't materialized.We want to emphasize that we're currently experiencing an extremely volatile GPU market characterized by limited supply, heavily inflated prices, and considerable issues and manipulation throughout the supply chain.Of course, we want to emphasize that we're currently experiencing an extremely volatile GPU market characterized by limited supply, heavily inflated prices, and considerable issues and manipulation throughout the supply chain.The GPU values presented here might look different in a few weeks, especially if supply improves and competition increases among GPUs in stock. Increased supply and market saturation are essential to driving down GPU prices, and there are plenty of reasons this might happen. Primarily, AMD needs to build market share by pumping out their in-demand cards and getting them into the hands of gamers.On top of this, it's important to remember that graphics cards typically sell out around launch time. This has happened many times before, and usually, the market begins to stabilize after a month or two. While this clearly hasn't happened yet with the RTX 5090, we're beginning to see early signs with other models.In Australia, for example, you can currently purchase an RTX 5080, RTX 5070 Ti, RTX 5070, or RX 9070 if you really want to. Yes, most available models right now are extremely overpriced special editions rather than base MSRP cards, but just a week or two ago, even these overpriced cards were unavailable. Once these overpriced GPUs start sitting unsold on shelves, mid-range models will follow, and eventually, reasonably priced GPUs should become available again.
