It’s no secret that Nvidia has been leading the charge in Generative AI with a huge surge in notoriety but also in profits becoming a wildly successful business overtaking the likes of Amazon and Alphabet (Google’s parent company). With that massive amount of success, however, it begs the question of where exactly a new generation of graphics cards comes in and if we even need one. As someone who’s backed Team Green for over 10 years, I’m leaning on the latter.
Here are the facts: it’s 2024 and we’re on track to receive a new graphics card generation following on from the RTX 40 series launch in Q3 2022. While the RTX 50 series has yet to be officially confirmed by Nvidia, that’s likely to be the direct continuation as it follows on from Ada (RTX 40 series), Ampere (RTX 30 series), and Turing (RTX 20 series) before it. Each cycle of two years is for a new architectural leap, but that’s without the caveat of the company’s unprecedented shift outside of gaming.
For those engrained in the PC gaming scene, Nvidia is the maker of some of the best graphics cards having carefully and steadily built up a reputation with its GeForce line of hardware for over 20 years. Although, as its reputation has exploded in the public consciousness in the last handful of years, if you ask the average person, they’re going to think of the company as an AI powerhouse. That spells contention for how a mega-corporation rethinks its position in the market, as the last GPUs hinted at.
The paradigm shifted with the RTX 40 series
Nvidia garnered favor with the RTX 30 series as the second-generation real-time ray tracing graphics cards proved to be head and shoulders above the rockier RTX 20 series. However, this GPU generation was marred by some serious stock issues, leading to scarcity, and overinflated prices through major retailers. Enter the RTX 40 series, which saw the potential for price increases and ran with it, and suddenly, an otherwise excellent selection of powerful hardware was victim to overpricing.
Suddenly, in two years, Nvidia came out of the gate charging a hefty markup on their 80-class video cards, evidenced by the RTX 3080 (selling at $699) to the RTX 4080 (retailing for an eye-watering $1,199). That’s a bitter pill to swallow, and both ourselves and other publications pulled this decision up in Nvidia RTX 4080 reviews. It’s something that Team Green itself attempted to course-correct earlier this year with the RTX 4080 Super by dropping MSRP by $200.
Essentially, Nvidia saw that consumers were willing to pay over the odds for its graphics cards through a difficult time for the industry financially and ran with it. This even led to complications around the RTX 4080 12GB which was eventually un-launched and then re-launched as the RTX 4070 Ti, itself viewed as overpriced and underpowered before the RTX 4070 Ti Super came in and beefed things up. The revised card was even built on the larger AD103 die with the same 16GB as the 80-class card, too. Something to be commended for sure. I fear that prices will surge again, but that’s only part of my worries.
Blackwell architecture has been revealed as an AI powerhouse
As touched upon above, Nvidia has been far more profitable and successful in the world of Generative AI innovations than playing nice with the gaming crowd. So it wasn’t surprising when Jensen Huang revealed the Blackwell B200 dual-die at the GTC March Keynote for the architecture’s usage in AI instead of gaming prowess. For two hours, this conference unveiled the die being 4x more powerful than Hopper architecture with no time to talk about any application for gamers.
From Disney Robots to Project Groot, the Omniverse, and various microservices all have to do with how the work will work with incredible amounts of data being transmitted through deep learning and large language models. That doesn’t leave much in the way of how we’re going to play games in the future. The most obvious throughline here is to look at how the die could further innovate on DLSS with a successor to Frame Generation, effectively eradicating native performance at the top end.
It’s been rumored that the RTX 5090 and RTX 5080 could be coming at the end of this year, which is on track with previous generations, but with the RTX 4090 already able to do native 4K at well above 60fps and even 8K at 30fps, where exactly does that leave a successor line to go? It opens a further dialogue that is hampered by something else entirely, console parity, as now the games we play will start to be restricted to what the Xbox Series X and PS5 are capable of doing with their 2020 tech.
That’s right. Despite games looking and running better on top-end PCs than consoles, the software is still made with the mainstream market in mind, and this extends to the hardware. The most recent Steam hardware survey shows that more gamers are using the older RTX 3060 and RTX 2060 graphics cards than their Ada equivalent. Broadly speaking, this is roughly about what the consoles can do, and therefore the standard for AAA developers and publishers. As time goes on, it’ll be even more of a bottleneck as this hardware is beginning to show its age.
This all culminates in uncertainty around the RTX 50 series
This culmination of factors; Nvidia’s wild financial success in the world of AI combined with the potential for further price increases and the stagnation of video game fidelity all leads me to be worried instead of excited about the RTX 50 series. Technically, it could be amazing, but if it costs even more and leans further into prowess around AI than gaming then it’s going to seriously alienate even more people.
The RTX 50 series could come out cheaper than its Ada counterparts with vastly improved performance and prove me wrong. However, the innovations and dominance the company has established in Generative AI means it doesn’t have to fight for its place in the market and can set its valuations. Whether this means the next generation is aimed more towards deep learning with gaming as a secondary concern or creates a new GPU class remains to be seen, either way, the signs are mixed right now.
Will there be an RTX 50 series?
While not confirmed yet by Nvidia all signs point toward there being an RTX 50 series reveal towards the end of the year.