Looking at the current GPU market, there has been almost no change in prices.

With the advent of the new generation of GPUs, and the upcoming new graphics cards from the three major chip manufacturers, we are about to face a moment where you have to decide whether to buy the graphics card you need immediately or wait for the next generation.

Based on some discussions we recently had at Computex, let's find out the most reasonable choice by looking at the current GPU market and expectations for the rest of the year. We have confirmed from several motherboard suppliers that among the upcoming graphics cards, only Nvidia will offer high-end options.

The decision to buy now or wait can be divided into different categories, depending on the type of GPU you usually buy. Do you mainly buy high-end graphics cards or cheap ones? Are you only considering Nvidia GPUs, or are you interested in all GPU brands? How old is your current graphics card? Is it performing poorly in games? Are there any upcoming games you want to play with the highest quality settings?

High-end GPU market and rumors of RTX 50 release

Let's start with high-end buyers. We have confirmed from several motherboard suppliers that in the upcoming generation of graphics cards, only Nvidia will offer high-end options. We are not sure what Intel will do, but we expect the Arc Battlemage to not have flagship-level performance.AMD's RDNA 4 series is currently only planned to launch mid-range models. Therefore, if you are pursuing high-end performance above $700, GeForce will be your best choice.

When speaking with numerous individuals familiar with the upcoming GPU, the specific release time of the GeForce 50 series is a matter of debate. However, the general expectation is that the first RTX 50 graphics card will be launched in the last few months of 2024, and it will be a high-end card.

The specific models or when they will be launched have not yet been determined, but the information we have heard generally falls into two categories: either the RTX 5080 and RTX 5090 will be launched closely together around the end of 2024, or the RTX 5080 will be launched at the end of 2024, followed by the RTX 5090 in early 2025.

We are discussing plans for 4 to 6 months from now, so nothing is certain, but motherboard suppliers are confident that these models will be launched first, with the 80-tier card expected to be launched before the 90-tier card.

With this in mind, for high-end buyers, we recommend waiting before upgrading your GPU. Most high-end GPUs are already 18 months old and are in the last few months of their release cycle, with prices having hardly changed in the past six months. If you were not attracted by the RTX 4080 Super launched at $1,000 in January — this was the last major release and price change in this segment, then this model will not attract you today, as its price is essentially the same.

The same is true for AMD's Radeon RX 7900 XTX, which has been priced between $900 and $950 for the past 12 months. The price of the RTX 4090 is still very expensive, around $1,750, and it is not worth paying more than the manufacturer's suggested retail price for a deck to be replaced at the end of the year.

We do not know whether the GeForce 50 series GPUs will significantly improve cost-effectiveness, but Nvidia is mainly competing with itself in this price range. If they cannot provide a significant leap in cost-effectiveness in the range above $700, they will not be able to attract existing users of the RTX 3080 for an upgrade.

The cost per frame of the RTX 4080 and RTX 4070 Ti is essentially the same as the RTX 3080, so it is unlikely that a third-generation product with similar value will sell in large quantities. Moreover, even if the value does not further increase, waiting for the launch of the 50 series will also provide you with the most comprehensive information to make the best purchase decision.Mid-range GPU Market and RDNA 4 Release Rumors

For buyers of mid-range graphics cards, waiting is also reasonable, albeit to a lesser extent. The next major mid-range graphics card is expected to be released in early 2025, namely AMD's RDNA 4 series. We have heard from several motherboard suppliers that AMD plans to launch its next-generation GPU at CES 2025 in January.

Since AMD is targeting the mid-range market with RDNA 4, the next-generation Radeon GPU should theoretically offer better value for money than existing graphics cards. We are not sure what the top performance level of RDNA 4 will be, but there are rumors that it will be comparable to the Radeon RX 7900 XT, which has been retailing for $700 since the beginning of this year. Purchasing the 7900 XT or an equivalent GeForce card like the RTX 4070 Ti Super now is not an ideal choice, as the price for this level of performance will drop by $100-200 in early next year.

However, we are also talking about something six months away, and a large number of new games will be released between now and then, such as "Star Wars: Rogue Squadron," "Assassin's Creed: Shadow," "Call of Duty: Black Ops 6," "God of War: Ragnarok," and "Dragon Age: The Warden's Veil." Depending on the graphics card you are currently using and your interest in the holiday 2024 game releases, it may be wise to upgrade sooner. Please note that the prices of current mid-range GPUs have not changed significantly this year, so the price of any graphics card you buy now is likely to be similar to the price in January.

In any case, given that prices have not changed much in the past few months, I would wait for the holiday GPU promotions in November. I believe that AMD, Nvidia, and motherboard suppliers will try to launch existing models at discounted prices during this period before the new GPU release, to avoid the situation of unsold backlog of previous models.

With AMD expected to launch the RDNA 4 GPU in January 2025, how will the mid-range GeForce develop? We have not heard much about these graphics cards, indicating that Nvidia does not plan to release them in the near future. Both AMD and Nvidia are extending the release cycle of each generation, so mainstream GeForce 50 graphics cards may take some time to emerge.

For example, going back to the GeForce 10 era eight years ago: the GTX 1080 was launched in May, followed by the GTX 1070 in June, and the GTX 1060 in July, entering the mainstream market in just two months.Regarding the 20 series, we obtained the RTX 2080 Ti and 2080 in September 2018, followed by the RTX 2070 in October, and the RTX 2060 in January 2019—spanning a four-month release period.

The 30 series was launched in September 2020 with the RTX 3080 and 3090, the RTX 3070 in October, the RTX 3060 Ti in December, and the RTX 3060 in February of the following year. It has now been five months since the launch.

Recently, the release pace of the RTX 40 series has been the longest to date. The RTX 4090 was introduced in October 2022, followed by the RTX 4080 in November, the RTX 4070 Ti in January 2023, the RTX 4070 in April, the RTX 4060 Ti in May, and finally the RTX 4060 in June—taking a full eight months for the mainstream models to hit the market.

Based on this, as well as the expected release times for high-end models, we will not see mid-range Nvidia 50 series GPUs until the latter part of the first quarter of 2025 at the earliest—if AMD launches the RDNA 4 first, this process could be accelerated. We just want to say that we should not expect mainstream RTX 5070-type products a few months after 2025, which could affect your decision to buy now or wait for more than nine months for new products.

Entry-level GPU market

For those who usually purchase mainstream to entry-level graphics cards in the market below $400, we expect new products in this range to be introduced no earlier than early to mid-2025. Particularly for graphics cards close to $250, it may not be until mid-2025, a year later. This is based on the release schedules of AMD and Nvidia's products from the past few generations, rather than any confirmed information.

The unknown here may be the new Intel Battlemage GPU, rumored to be launched at any time between the end of 2024 and the beginning of 2025, possibly slightly earlier than AMD and Nvidia's products.

If your current graphics card is not sufficient and you want to play the latest PC games, waiting a whole year to upgrade is not a good suggestion. At the same time, the current GPU value is not astonishing, and the prices have been stable for a while as each company has released its full range of products.The Radeon RX 7800 XT and RX 7700 XT are the only two GPUs that have seen a significant price drop since the beginning of 2024. The competitor, GeForce RTX 4070, has been consistently priced at around $550 since last September, while the RTX 4060 Ti 16GB reached its current price of $430 in November 2023.

Summary

Looking at the current GPU market, there has been little change in prices by 2024. Base prices have dropped by 5% here or there, but most models have not seen significant discounts or price adjustments.

The largest increase is the Radeon RX 7700 XT, which is currently priced at $380. Nvidia released the Super series earlier this year, but apart from that, the market has been flat and stable.

For buyers, there are pros and cons. On the positive side, we are not in a period where you bought a graphics card a month ago, only to find that the GPU has become much cheaper the next month, which can be frustrating. It doesn't matter when you buy a graphics card; their prices have been stable for a while. The bad news is that consumers are not seeing a downward trend, so for a GPU market that was not reasonably priced from the start, GPUs will not become cheaper over time or return to a more "acceptable" level.

Nevertheless, the market looks healthier before the arrival of the new generation of graphics cards. When Nvidia and AMD were preparing to release the GeForce 40 series and Radeon 7000 series, there were still a large number of old graphics cards on the store shelves due to the surge in sales caused by cryptocurrency mining and lockdowns.

This will affect the decision-making of Nvidia and AMD in terms of graphics card pricing and release strategies. If either company is too aggressive in pricing, the large number of old model products that they and their partners are trying to sell will depreciate. And with so much remaining inventory, flash sales and discounts are not enough to provide incentives. This means that GPU suppliers must pay more attention to the pricing and inventory of the previous generation products than usual.

In 2024, the backlog of supply will not be so serious. The situation looks more "normal" – just like before the recent boom. An example is that after the launch of the 4070 Ti Super, Nvidia was able to easily clear the RTX 4070 Ti from the market – which was impossible during the RTX 30 series era. The inventory of RDNA 2 is rapidly decreasing, with many cards no longer available, while the supply of RDNA 3 looks normal.

We are not saying that Nvidia and AMD will suddenly increase the value of the next generation of graphics cards because of this. But internally, there will no longer be a need to consider the existing models.The primary consideration for the next generation of GPUs will be other more profitable products that will deplete wafer allocations, such as Nvidia's AI data center GPUs and AMD's Epyc server CPUs.