Table of Contents
AI is purportedly good at providing information and insights when asked questions about myriad topics. But can it answer why we’ve heard very little about their nextits RTX 60-series? Regardless of where Nvidia goes next with consumer GPU naming and solutions, why hasn’t it confirmed that it’s moving on from the current Blackwell lineup? It’s been over a year, and…nothing. We wondered what ChatGPT, Perplexity, and Gemini had to offer. And no, it’s not all down to RAM.
The prompt: “Why do you think Nvidia has stopped its graphics cards at the RTX 5090?”
For a broad range of insights, we asked this across the three tools mentioned above. There was crossover, but below are the areas where each provided the most interesting considerations. This is a reminder that if you use a singular AI model for your prompts, others may get slightly different responses elsewhere. And also, remember that AI can make mistakes.
Latest PC & Tech deals
Acer Predator Helios Neo 16 - was $2,199 now $1,749
ASUS SFF-Ready RTX 5080 - was $1,499 now $1,354
Samsung Odyssey G7- was $999 now $649
HP OMEN 45L gaming PC - was $2,499 now $2,309
BIWIN Black Opal NV7400 4TB SSD - was $599 now $459
Prices correct as of July 10th, 2026.
ChatGPT – ‘There isn’t much room above it’ #
There isn’t much room above it.
If NVIDIA released an RTX 5090 Ti or RTX 6090-class product in the same generation, it would need to be:
- noticeably faster
- stable at reasonable power
- and worth the extra cost.
The RTX 5090 already draws a great deal of power and uses one of NVIDIA’s largest gaming GPU dies. Any additional performance would likely require disproportionate increases in power consumption and cooling, making it difficult to justify as a consumer product.
Yes, correct, this is an issue. How many gamers do you know who have an RTX 5090? While it is the most powerful, it’s not the best graphics card for gaming, and the Steam Hardware Survey (June) shows that the card represents a minor 0.41% of Steam users’ systems. In comparison, the far more modest RTX 4060 has 3.46%. The clear fact is that the 5090 is still far too powerful for most.
It’s also so powerful that any ‘6000-Series’ counterpart may struggle to offer clear benefits above it. So, business-wise, any RTX 6000 release would seem to make little sense right now. Further, noting that “Nvidia makes far more money from its professional lineup than from gaming GPUs”, ChatGPT suggests there is “little incentive” for the company, as RTX Pro workstation cards, data-centre GPUs, and AI accelerators “generate much higher margins”.
Perplexity – ‘Nvidia likely hasn’t “stopped”’ #
My read:
So the short version is: Nvidia likely hasn’t “stopped” so much as chosen the RTX 5090 as the current ceiling for consumer GeForce, while directing its biggest engineering and business energy toward AI/datacenter chips and possibly future refreshes. The rumors about the 5090 disappearing appear to be more about inventory, regional availability, and speculation about a Super refresh than an actual end of the line.
I’m not sure Perplexity quite understands the brief here, but perhaps the prompt wasn’t clear. The RTX 5090 disappearing isn’t what I am interested in, and yet the AI tool suggests rumors about that while pointing out speculation about a ‘Super’ refresh. Yet there’s no mention of a ‘6090’ or even a future 60-Series GPU generation.
Yet it does back ChatGPT’s insight on business strategy, focusing on “energy towards AI/datacenter chips”. That seems fairly likely because, according to Nvidia’s quarterly earnings, its Edge Computing is the least profitable platform for the company. Edge computing includes its consumer graphics card products, and it massively lags behind the likes of Data Center and AI
Gemini – ‘It actually hasn’t!’ #
It actually hasn’t! The idea that Nvidia “stopped” at the RTX 5090 comes down to a slight misunderstanding of how they name their graphics cards. The GeForce RTX 5090 is simply the premier flagship of their current RTX 50-series (built on the “Blackwell” architecture).
It’s cute of Gemini to consider that I’m misunderstanding what’s going on, but it’s filling in the gaps, and this confusing response is likely a prompt issue. So, it’s a good reminder to be clear and consider the permutations when using AI for knowledge hunting. Still, Gemini also surfaces the shift towards AI and data centers (there’s something to it) and suggests another view as well.
We are hitting the physical limits of how small transistors can be made on a single piece of silicon. The RTX 5090 is already a monster that draws 575 watts of power natively. Making a “bigger” traditional gaming card right now is incredibly difficult without it requiring its own dedicated power outlet or being too expensive to manufacture. Instead of pushing past the 5090 tier, Nvidia has been rumored to fill structural gaps with cards like an “RTX 5090 SE” or “Super” mid-generation variants.
This is a technical consideration that few, except informed or expert enthusiasts, would consider. So Gemini has dug up a bit of a gem here to be polished and investigated further. There are indeed limits to the number of transistors on silicon. Additionally, an Nvidia RTX 5090 SE or Super is possible, and there are rumors that an RTX 5090 SE could be coming.
Sadly, we cannot know when we’ll see a new graphics card beyond the RTX 5090, in power or generation. But while AI is incapable of predicting the future, I’m very surprised that no tool picked up on the cost/availability of memory chips, RAM, or VRAM as a reason why Team Green is all quiet on what’s next. Maybe that cost is more of a consumer problem. Or maybe the bigger, honest issue is that Nvidia simply has other priorities.