Three new RTX 40 SUPER-Series GPUs will be introduced as two other cards are phased out, greatly improving the value proposition of NVIDIA’s gaming lineup.
Three new RTX 40 SUPER-Series GPUs will be introduced as two other cards are phased out, greatly improving the value proposition of NVIDIA’s gaming lineup.
NVIDIA underscored their dominant position in the GPU market today with their CES 2024 livestream, revealing this winter’s worst-kept secret in IT alongside in a relatively relaxed presentation highlighting leadership in gaming and AI. Three new SKUs in the RTX 40-series – reviving the SUPER label for refreshed cards – revised their price/performance landscape, new RTX-supporting titles were announced, and AI was touched on with a certain amount of well-warranted smugness; but with little to excite consumers was it enough for the viewers at home?
Rumours of a GeForce RTX 40-series refresh in the form of SUPER series SKUs have been doing the rounds for months, with only final specs and pricing seemingly up for debate. Emphasis has naturally been placed on where these cards will sit on price/performance curves, and so it comes as something of a relied for the green team finally place some much-needed certainty in the minds of prospective buyers.
Joining the RTX 40-series this month will be the GeForce RTX 4080 SUPER, RTX 4070 Ti SUPER and RTX 4070 SUPER, replacing both the RTX 4080 and RTX 4070 Ti which are now in the process of being phased out. Each are a modest bump in performance over the SKUs they replace, but the more significant aspect of their release is their targeted price points.
Launching on January 31st, the RTX 4080 SUPER is equipped with a full-fat AD103 GPU boasting 10240 CUDA cores, 16GB GDDR6X memory and a modest improvement in operating frequency over the GPU it replaces. It’s also coming in at an MSRP of $999, fully $300 less than the RTX 4080 at launch and just enough to ruffle the features of those who might have purchased the card in recent months. Performance is anticipated to be roughly 5% faster than the RTX 4080, and significantly faster than the RTX 3090 Ti flagship from the previous generation.
Coming in at $799 on January 24th is the RTX 4070 Ti SUPER. It features a healthy 768 core bump in shader count to 8448 and slight improvement to base clock speeds, but potentially the biggest difference for real-world performance might be its increase in VRAM from 12GB to 16GB of GDDR6X memory over a wider 256-bit bus. It should mean far better performance at higher resolutions and image quality settings compared to the non-SUPER variant, making it effectively a cut-down RTX 4080 rather than maximised RTX 4070.
Finally, and arriving in only a week, is the $599 RTX 4070 SUPER. Rather than replace the RTX 4070 it will sit just above it in the product stack, potentially causing downward pressure on the RTX 4070’s price going forward. It boasts over 1000 more CUDA cores but otherwise maintains the same specs as its predecessor, and in theory should outperform an RTX 3090 even before 40-series-exclusive features such as Frame Generation are taken into account.
It’s expected that the new SUPER-series cards will all feature new 12V-2×6 rather than the now defunct 12VHPWR connectors, and each will probably be shipped with an appropriate adapter for PSUs which don’t support the standard natively. UK Pricing inclusive of V.A.T. is expected to come in at £959, £769 and £579 respectively. A full complement of partner cards will be part of each SKUs launch, offering proprietary features, cooling solutions and aesthetics according to gamer preferences.
This refresh makes the RTX 40-series far more competitive overall against AMD’s relatively strong equivalents, but NVIDIA’s dominance isn’t just about its hardware. Also part of the presentation was the announcement of new RTX titles which support NVIDIA-exclusive features, joining a list now more than 500 strong. Among them are Diablo 4 – finally receiving Ray Tracing support in March – Horizon Forbidden West – soon to boast DLSS and DLAA when it comes to PC later this year – and social sandbox MMO Pax Dei – which will enter Early Access with DLSS 3 support.
Also receiving some love is GeForce NOW, NVIDIA’s game streaming platform. In addition to new titles, particularly from previously sceptical publisher Activision Blizzard, the service will soon support G-SYNC Technologies to improve stream quality and reduce latency. Reflex modes will receive 60fps and 120fps modes, boosting GFN’s competitive gaming credentials. And a new Game Day Pass system will allow free members to experience the very best of GFN servers, before committing to a longer-term subscription.
Last but not least, NVIDIA also devoted some time to the potential new generative AI techniques have in the field of gaming beyond graphical fidelity. The most exciting possibility might be AI-generated an managed NPCs, with dynamic speech and movement for a far more immersive storytelling experience. The system already appears much more fleshed out than 6 months ago, but the real litmus test will be in seeing what talented game developers can do with the cloud-based tools NVIDIA will be offering.
There’s plenty more in the field of AI and industry to be seen in the presentation as a whole, which you can catch below.
