Nvidia last released a graphics card in March of this year, and it is currently the strongest GPU in the market for gaming. The RTX 3090 Ti has the best specifications of any gaming GPU, but it costs as much as a full-fledged average gaming setup. The $2,000 price tag has prevented a lot of gamers from buying it.
Currently, there is no point in buying the Nvidia flagship graphics card because it's more sensible to wait for a few months and buy the next generation of GPUs that will perform better and cost less, a pattern observed with every new generation of technology.
How much better will the new Nvidia cards perform?
The new 40 series cards by Nvidia are codenamed "Ada Lovelace" and are rumored to be released in Q3 of this year, which can be any time from July to September.
The usual cards can be expected this year: the RTX 4060, RTX 4070, RTX 4080, and RTX 4090, and the refreshed Ti versions will likely be released next year. The weaker RTX 4050 can be expected this year too, but it is not certain. This is because the RTX 3050 was not released with the bunch in 2020, and it was only released in January this year.
Performance and design
We can expect the new cards to double the performance of the 30-series cards in synthetic benchmarks, which does not necessarily translate to real-world performance like video games and resource-intensive software, but it is still a significant boost. This can be said because Nvidia's 30-series offered a significant jump over the 20-series cards, and the same can be expected this time.
The RTX 40 series cards are expected to use the latest VRAM technology of GDDR6X. This technology has previously been seen in only RTX 3080 and RTX 3090 varients and some Ti versions of RTX 30 series cards. With some more improvements seen in this technology, the maximum amount of VRAM loaded in the RTX 4090 Ti could be up to 48 GB, which is overkill, but a possibility nonetheless.
The 40 series cards are expected to be massive in size with a bigger baseplate and a huge heatsink. But to power such a monster, it would not be a surprise if the wattage requirements skyrocket, with the TDP of the highest-tier GPUs being up to 600W, which will require a PSU that has a minimum power of 1200W.
Competition from AMD
AMD recently announced their RDNA 3 architecture that will be used by the RX 7000 series GPUs scheduled to release this year as well. In an interview with Rick Bergman, AMD's executive vice president, it was stated that AMD would be using an advanced manufacturing process that will deliver performance-per-watt improvements and make a leap of 50% in performance.
A similar improvement can be expected with Nvidia, or even more, as they have generally one-upped their competition in terms of performance.
Price
It's hard to say how much the next generation of graphics cards might cost, but rumors suggest that with efficient manufacturing by Nvidia, the GPU costs could be lowered. Taking the prices of the RTX 30 series GPUs as a guide, the prices can be predicted. The lowest tier 40 series card should start at around $349, with the mid-tier being around $499, and the top ones touching $1,499. Next year, the improved Ti cards are predicted to launch with their prices costing $100 more than their counterparts.
The prices of the current generation of GPUs are still not settled due to the global silicon shortage paired with scalpers buying up all the stock. So there is a high chance that next-gen GPUs will not be sold at MSRP, but at a higher price by resellers.