Nvidia 4060 Ti 8GB offers up the sort of performance Nvidia promises – and questions the current GPU market
With the GeForce RTX 4060 Ti 8GB, Nvidia has a clear audience and a clear goal in mind. We all like to make a big song and dance about the hardest-hitting, heaviest-lifting state-of-the-art graphics cards – but the truth is, the majority of people don’t buy them.
Glance over at the Steam Hardware survey and you’ll see that story: right now, the most popular GPU on Steam is still the Nvidia GeForce GTX 1650, followed by the 1060. Of all of the cards from the previous two generations, the chart-toppers are the RTX 3060 and 2060. A pattern is emerging, y’see? Those entry-level cards matter.
Nvidia’s own data backs this up. The company collects data on how people use their cards and software, and it reckons that 77% of players on PC are still running at 1080p or less. While on console 4K is becoming more ubiquitous, since when you buy a new living room TV you tend to buy high-spec, it can be hard to justify an expensive 4K & Gsync compatible display for your PC gaming goblin den. The numbers back it all up.
In that sense, this makes the 4060 Ti one of Nvidia’s most important cards of this hardware generation – and it’s designed to directly provide an upgrade path to those on similar cards of older generations. There’s three variants, with the 4060 Ti 8GB out now for $399, to be followed by a 16GB variation for around $499 in July. Finally, we’ll also get a non-Ti 4060, also likely in July.
I’ve been tooling around with the 8GB version of Nvidia’s Founders Edition 4060 Ti for a little over a week now, and my feelings about it are… complicated. In some ways, this card continues to be a wonder that demonstrates the real strengths of this generation of GPUs. In others, it’s a disappointment and even a bit of a slap in the face to consumers.
The single most impressive thing about the 4060 Ti is something it has in common with the rest of the 40 series, and something I’ve stressed in pretty much every Nvidia GPU write-up this generation – its Power Consumption. 40 series cards are alarmingly more efficient than their 30-series predecessors, and that continues here. I actually think this element of the GPU arms race gets understated, but in a cost of living crisis where running powerful gaming hardware can unsuspectingly get expensive over time, making one of the most power-hungry pieces of a PC more efficient is a huge boon – and such is the case here.
The good news continues whenever Nvidia’s built-in and exclusive technology is invoked. DLSS is a game-changing technology for video games, and it’s no longer theoretical or a nice idea – it’s real, here to stay, and in my opinion is a must-have. If a game has it available, you turn it on.
On weaker cards, the ability of DLSS is arguably more important. DLSS uses AI and the unique architecture of the RTX cards to render games at a lower resolution and then upscale them to be virtually indistinguishable from native higher resolution rendering. On a 4090, this helps you to push a 4K game with ray tracing on well past 60fps; maybe even to 120fps. But on an entry-level card like the 4060 Ti the goal is more modest – like getting a cutting-edge game to run at a rock solid frame rate even with ray tracing significantly increasing the render load.
Basically, the value proposition of DLSS shifts at this level. It’s not about pushing past the red line as much as it is keeping things nicely ticking over, reliable and manageable.
Case in point, I tested Cyberpunk 2077’s ray tracing getup, including RT Overdrive, with the 4060 Ti. Such a thing would’ve been unthinkable in previous hardware generations – but on this card, it’s possible. With DLSS 3 enabled on my favored settings (which includes Ultra RT), it’ll offer a great experience well in excess of 60fps at 1080p, and gets perilously close to 60fps at 1440p. For comparison’s sake, this is a very tidy uplift on previous generations: over 75% above the RTX 2060 Super, and a more modest 10-15% uplift on its direct predecessor, the 3060 Ti.
Cyberpunk is a banner example, probably the biggest graphical powerhouse on PC right now, this generation’s Crysis. But you can repeat this sort of testing with other DLSS3-enabled VG247 test favorites such as Forza Horizon 5 and Hitman, and the results are generally similar. But when you get to raw, unassisted rendering – basic rasterization – this card begins to struggle to justify its place a little more.
Here’s the brass tacks: at 1080, you’ll enjoy a modest uplift; in most games between 10% and 20%. Sometimes you can get a little lower, sometimes higher. But it’s all just… a little flat. Since the RTX series made its debut with the 20 series, we’ve been used to impressive generational leaps – and this just isn’t really that. There’s no ‘wow’ factor here, it’s belt-and-braces stuff, as far as an upgrade goes.
Even in RTX poster child Cyberpunk, the story echoes; it’s a strong performance for the price bracket and a decent way to play a demanding game with the 40 series-exclusive DLSS 3, but turn that setting off and you’ll see a generational leap over the 3060 Ti of under 10%.
And all of this brings us, of course, to the price. Nvidia is keen to point out that, generationally, this hasn’t had a price increase. It also hasn’t had the performance increase you’d expect across the board given it’s a generational shift, however. At a lower price point – or perhaps at this price point but with 16GB memory instead of this 8GB – this card would be a much easier recommendation.
As it stands, consideration of this purchase has to be deeply caveated. How many games that you want to play support DLSS? And perhaps more importantly, can you afford to wait out for a new generation, or to save up for one of the more impressive cards that also supports that technology? I can see the market, and I can see the interest – but I practically do so through deeply gritted teeth.
More than anything, the RTX 4060 Ti and the kicking it’s taken in the media this last week underline where the current GPU market is at. AMD isn’t really competing, and Nvidia, high on AI buzz and with a soaring company value, doesn’t appear too concerned about blazing a trail. That leads to a card like this: functional, but flat, and ultimately unexciting. Hopefully the base 4060 provides a better value.