Okay, so Nvidia kicked off CES 2025 by doing what they always do: reminding us that they basically invented gaming as we know it. You know, the classic “started from the bottom, now we’re here” story, complete with arcade machines and AI that can probably make chai by now. But let’s be real, we’re here for one thing and one thing only: shiny new GeForce GPUs. And boy, did they deliver, with the Blackwell architecture finally making its grand debut.
First up, the RTX 5070. At ₹59,000, it supposedly packs the power of an RTX 4090 at a third of the price. Yeah, we’ve heard that one before. Remember the 4000 series launch? “4070 performance at 4080 price!” Let’s just say we’ll believe it when we see it. But hey, if it actually delivers 4090 performance in laptops while sipping power like a hummingbird, that’s something to get excited about.
Now, for the rest of the lineup. We’ve got the RTX 5090, the big daddy of them all, priced at a cool ₹2,14,000. Then there’s the RTX 5080 at ₹1,07,000 which seems to be the sweet spot in terms of price-to-performance (unless they pull a 4080 again and launch a “Super” version a month later). And of course, the RTX 5070 Ti at ₹80,000, because why not have another tier?
But hold on a minute. Before we all rush to empty our bank accounts (or sell a kidney), let’s talk about those specs. Nvidia loves to throw around big numbers like “3,400 AI TOPS” and “125 TFLOPS,” but what does it all mean? Well, it seems like they’ve doubled down on AI performance, which is great if you’re into, you know, training your own robot army. But for gaming? We’ll have to wait and see.
And speaking of waiting, we’re still missing some crucial details about these cards. Memory speeds, transistor counts, die sizes – it’s all a bit of a mystery. Maybe they’re trying to build suspense? Or maybe they just haven’t figured it all out themselves.
One thing that did catch our eye is the VRAM situation. The RTX 5090 gets a whopping 32GB, but the rest of the lineup is stuck with 16GB or even 12GB. Now, that might be fine for today’s games, but with the way things are going, we wouldn’t be surprised if we need 32GB just to run Minesweeper in a few years.
But hey, at least they’re finally talking about AI-based texture compression. This could be a game-changer (pun intended), potentially allowing us to squeeze more performance out of our GPUs without needing a tanker truck full of VRAM. Of course, there’s the small issue of cross-platform compatibility. Will developers bother to support Nvidia’s fancy new tech if it means leaving console gamers behind? Only time will tell.
Speaking of fancy new tech, Nvidia also unveiled some seriously impressive AI-powered rendering techniques. They’re introducing something called “RTX Neural Shaders,” which basically lets AI take over the job of creating realistic materials and lighting effects. Imagine games with graphics so lifelike, you could practically reach out and touch them (though we wouldn’t recommend trying to touch your screen).
And get this: they’ve even got AI working on rendering human faces in real-time. With “RTX Neural Faces,” they can take a basic 3D model and turn it into a hyperrealistic digital human, complete with all the subtle nuances of facial expression. Combine that with their new ray-tracing tech for hair and skin, and you’ve got virtual characters that look almost indistinguishable from real people. Move over, Shah Rukh Khan, there’s a new king of Bollywood in town! We witnessed a demo of this technology a while back and spoke about it.
Of course, all this AI magic comes at a cost. Literally. Those RTX 50-series GPUs aren’t going to be cheap, especially once they hit the Indian market. But hey, if you want the best of the best, you gotta be prepared to pay for it, right?
In the meantime, we’ll just have to keep drooling over those DLSS 4 demos and dream of a world where our games look like Pixar movies and run at a million frames per second.
