Why I’m feeling hopeful about Nvidia’s RTX 50-series GPUs

Trending 2 weeks ago
The RTX 4070 Super connected a pinkish background.Jacob Roach / Digital Trends

I won’t dishonesty — I was beautiful scared of Nvidia’s RTX 50-series, and I guidelines by nan sentiment that those fears were valid. They didn’t travel retired of bladed air; they were fueled by Nvidia’s attack to GPU pricing and worth for nan money.

However, nan RTX 40 Super refresh is simply a measurement successful nan correct direction, and it’s 1 I ne'er expected to happen. Nvidia’s astir caller choices show that it whitethorn person learned an important lesson, and that’s bully news for early generations of graphics cards.

The value of performance

Nvidia really didn’t clasp backmost successful nan RTX 40 series. It introduced immoderate of nan best graphics cards we’ve seen successful a while, but earthy capacity isn’t nan only point to see erstwhile estimating nan worth of a GPU. The value is nan 2nd awesome facet and weighing it against capacity tin often extremity nan scales from “great” to “disappointing.” That was nan lawsuit pinch respective GPUs successful nan Ada generation.

Let’s return nan RTX 4070 Ti, for instance. The GPU could (or moreover should) person had much representation and possibly much CUDA cores, but those problems wouldn’t person been arsenic overmuch of an rumor if it was cheaper. With a $800 value tag, it was excessively costly for nan benignant of capacity it was capable to provide.

The RTX 4080 was successful nan same, aliases moreover worse, position. A powerful GPU successful its ain right, tin of tackling beautiful overmuch each crippled astatine 4K, it was doomed from nan commencement and dubbed nan worst-value GPU of nan year. It was priced astatine $1,200, truthful that sentiment was wholly valid.

PNY RTX 4080 pinch nan powerfulness connector attached.Jacob Roach / Digital Trends

Things don’t get amended arsenic we move down nan merchandise stack. The RTX 4070 was excellent, but each 3 versions of nan RTX 4060 were a letdown. Performance, VRAM, and questionable pricing each came together to create 3 reasonably “meh” GPU releases.

Atop nan lineup sits nan awesome RTX 4090. While it was outrageously costly astatine launch, it’s moreover pricier now, but considering nan generational leap betwixt nan RTX 4090 and its predecessor, that value made a batch much consciousness than immoderate of nan different choices Nvidia made this generation.

With a lineup of awesome GPUs wherever each had thing that could usage improvement, Nvidia had a batch to dress up for. Personally, I felt disillusioned and didn’t judge that it would return immoderate steps to hole nan sorry authorities of nan RTX 40-series.

And then, Nvidia decided to beryllium maine wrong.

How Nvidia says ‘sorry’

The RTX 4080 Super graphics paper sitting connected a pinkish background.Jacob Roach / Digital Trends

The RTX 40 Super refresh added 3 much GPUs to nan lineup: nan RTX 4070 Super, nan RTX 4070 Ti Super, and nan RTX 4080 Super. Early rumors revealed nan specs of each GPU measurement up of time, truthful technically, we knew what was coming.

The RTX 4070 Super received a important boost successful CUDA halfway counts. The RTX 4070 Ti Super, connected nan different hand, addressed nan awesome problem galore group had pinch nan guidelines exemplary — a deficiency of VRAM. Now, nan Super type serves up 16GB of VRAM and an due representation interface. Coming successful last, nan RTX 4080 was nan large enigma anterior to nan release. With a hardly visible spec upgrade, nan GPU was destined to beryllium a sidegrade and a clone of nan RTX 4080.

Again, capacity is 1 thing. The prices remained concealed until correct earlier nan launch, and I afloat expected Nvidia to value each of these cards higher, conscionable connected nan ground of them being “Super.”

Boy, was I successful for a shock.

RTX 4070 Super RTX 4070 Ti Super RTX 4080 Super
CUDA cores 7,168 8,448 10,240
Boost clock 2,475MHz 2,610MHz 2,505MHz
Memory speed 21 Gbps 21 Gbps 23 Gbps
Memory bus 192-bit 256-bit 256-bit
TDP 220W 290W 320W
Price $600 $800 $1,000

Instead of yet different lawsuit of Nvidia ignoring nan important facet of performance-per-dollar, we sewage 2 cards pinch a noticeable bump successful performance, and nan RTX 4080 Super pinch a shocking value cut. That was all that it needed to go a awesome GPU.

For once, nan outcry of consumers that was heard crossed various societal media platforms was listened to. Nvidia addressed nan main issues we’ve had pinch nan RTX 4070 Ti (poor capacity for nan money and debased VRAM) and nan RTX 4080 (terrible value). The RTX 4070 Super was awesome to statesman with, and now, it’s moreover better, each astatine nan aforesaid price.

This is simply a uncommon lawsuit of a “win-win” crossed nan board, and while nan cards could guidelines to beryllium moreover cheaper, it’s important to admit that Nvidia consciously strayed from nan words of its CEO, Jensen Huang, who said that falling spot prices are “a communicative of nan past.”

This is simply a large deal, but it’s besides not nan first clip that Nvidia decided to opportunity “sorry” by fixing its past mistakes successful nan very aforesaid procreation of GPUs.

Not nan first Super

Nvidia's RTX 2080 and RTX 2080 Ti sitting connected a table.Riley Young / Digital Trends

Nvidia’s nary alien to having to apologize and hole mistakes that didn’t spell complete good pinch its customer base. The RTX 20-series Super refresh was akin to what we sewage successful this generation.

Months aft nan motorboat of nan RTX 2060, RTX 2070, and nan RTX 2080, Nvidia turned astir and decided to amended what went wrong. The Super versions came successful astatine nan aforesaid (or somewhat higher) value points, but pinch an betterment successful specs. Sounds familiar, right?

However, nan specification bump wasn’t monolithic capable to make overmuch of a splash. The RTX 2080 Super was only a fewer percent faster than its predecessor, while nan 2060 Super and nan 2070 Super saw bigger gains successful nan ballpark of 10% to 15%. Of course, their prices were little to statesman pinch than what we’re dealing pinch these days; an RTX 4080 priced astatine $700 would person been a slayer GPU.

On nan surface, not overmuch is different betwixt nan RTX 20-series refresh and nan 1 we’re watching this clip around. The older GPUs will mostly beryllium phased retired successful favour of their superpowered counterparts, and nan marketplace is amended disconnected pinch much GPUs up for grabs while some types are still up for sale. The motorboat of nan refresh was timely, giving Nvidia thing caller to talk astir earlier it’s fresh to motorboat nan RTX 50 series.

However, this clip around, nan refresh feels significant. Given its pricing strategy for nan RTX 30-series and nan RTX 40-series, Nvidia could person conscionable done nan aforesaid point and slapped a higher value connected each GPU. Sure, galore group would beryllium disappointed, but hey, it’s Nvidia — group would still bargain it. (Case successful point: myself. I bought nan RTX 4080 past year.)

We could person gotten nan RTX 4070 Ti Super pinch much CUDAs, but nan aforesaid VRAM astatine $900, aliases nan RTX 4080 Super pinch hardly immoderate improvements and a $1,300 value tag. We didn’t.

What does this mean for Nvidia’s next-gen cards?

The GPU scenery is changing

Two RTX 4070 Ti Super graphics cards sitting adjacent to each other.Jacob Roach / Digital Trends

For nan first clip successful months, I’m emotion hopeful. After all, it was ne'er nan rumored capacity of nan Blackwell procreation that worried me; it was ever nan truth that I didn’t judge Nvidia still cared to perceive to feedback. However, nan RTX 40 bid Super refresh tells maine otherwise.

Could it beryllium a one-off? Maybe. But it’s difficult not to announcement that nan user marketplace whitethorn slow beryllium reaching nan limit of really acold astir group are consenting to spell erstwhile building a caller PC. The prices tin only spell up truthful precocious earlier group move to cheaper options, beryllium it entry-level GPUs, consoles, aliases portables for illustration nan Asus ROG Ally. The second wouldn’t beryllium a awesome solution for Nvidia, seeing arsenic it still hasn’t dipped its fingers into nan cookware of golden that nan budding gaming handheld marketplace appears to be.

Now that Nvidia has shown america that it’s possible, I’m hoping that nan RTX 50 bid will grow connected that. Give america somewhat amended worth and GPU configurations that make consciousness — so, not thing for illustration nan RTX 4060 Ti pinch 16GB of VRAM and a 128-bit bus. Instead, fto america get an RTX 5070 that starts astatine $600 and outpaces nan RTX 4070 Ti, aliases an RTX 5080 that either boosts nan capacity aliases knocks down nan price. Maybe down to $900? One tin dream, but realistically, I’ll beryllium happy if it stays astatine $1,000.

Without a crystal ball, it’s intolerable to cognize what Nvidia will do next. It mightiness do nan aforesaid point arsenic nan erstwhile clip it released a mid-gen refresh — revert backmost to its aged ways and put our wallets done nan wringer pinch overpriced GPUs. That attack would, undoubtedly, person AMD rubbing its hands together successful anticipation of immoderate easy profits successful the midrange segment wherever gamers are likelier to take worth complete performance.

The point is, Nvidia can’t support doing nan aforesaid point everlastingly and hoping for a different outcome. The reception of nan RTX 40-series showed nan tech elephantine that conscionable slapping a Nvidia logo onto a merchandise won’t ever beryllium capable for it to sell. Let’s dream that it’ll transportation that instruction connected to nan adjacent procreation and present amended worth pinch nan RTX 50 bid successful 2025.

Editors' Recommendations

  • CableMod’s adapters damaged up to $74K worthy of Nvidia GPUs
  • Nvidia’s RTX Video tin upscale blurry YouTube videos
  • These early RTX 4080 Super listings confirmed my worries astir nan GPU
  • I’m worried astir nan Nvidia RTX 4080 Super
  • At this point, we cognize conscionable astir everything astir Nvidia’s caller GPUs