$40 cheaper than the previous deal.
Don't forget to enter in the discount code at checkout.
Card should also be eligible for a free copy of Doom.
Surcharges: 0% for bank deposit, Afterpay & Zip Money. 1.2% for VISA / MasterCard & PayPal. 2.2% for AmEx.
Free shipping excludes WA, NT & remote areas.
For 5080 Super 24gb announcement at computex next week?
5070 Super 18GB for me please.
computex is 20th May. it is likely $2200+, nearly double of the $1139 9070XT. it is out of my list now. I am hoping for a 24GB 9080 XT
That would definitely be a surprise drop by AMD.
Arc B770 16gb might also be a surprise drop later this year.
FYI the "origins" of that "rumour" is just one random forum guy with zero affiliations to the industry.
Super will likely launch 12 months after the original 5000 series cards.
damn thats a good price for a 5070 TI mine costed my 1800 bucks on launch so thats a very good price.
We'll see $1139 9070 XT again soon
I am under the impression nVidia still has the better drivers, proprietary technology, and the stronger cards… has this changed?
better drivers
Lol no the RTX 5000 series has literally been plagued with driver problems so this issue isn't exclusive to AMD
proprietary technology
Gaming wise, AMD and NVIDIA offerings are very similar. DLSS4 being slightly ahead FSR4. Productivity and AI stuff is a different story however, you'll want an NVIDIA card
stronger cards
At the top end, yeah (the 5090), but AMD offers much better value frame/dollar wise.
AFMF still needs the AI steroids to compete with DLSS framegen, let alone MFG.
Even though they’re derided by “snobby” reviewers, they are an extra feature consumers will consider over AMD.
interesting - thanks for opinion/perspective :)
Well the drivers seem to have gone downhill with the 50 series launch
AMD drivers have been mint for the past couple generations now. It's NVIDIA which has numerous driver issues now.
Really though? I thought that was just an AMD rebate thing
Correct - the initial launch pricing was subsidised by AMD. Not impossible for it to get back down to launch RRP, but it would require it to technically be discounted from current RRP likely as a result of excess supply (which seems optimistic in the current climate).
This whole GPU launch cycle has been horrendous from AMD and nVidia as well as from a product/price point of view and the obfuscation/misdirection in marketing.
The performance/dollar metric has stagnated or gone backward(!), and the only cards that provide performance you couldn't have already bought 2+ years ago are the 5090 and (kind of…barely) the 5080. Even now you can buy a 7900XTX with 24Gb of ram for around the price of the 5070Ti, or marginally more than the 9070XT…but you could've bought it ages ago sooo…progress?
With a free copy of doom?
Great price for the best value Nvidia card out there right now
As Steve Burke says, PC gaming is now for the rich.
stuff that.. 2ndhand 1070 still going strong ~7 years later
The 9070xt still $300 cheaper hahaha
Cheapest in stock is $1249, no? Would take a 5070ti over a 9070xt any day at $160 difference.
Have to say for $160 difference and doom dark ages game 5070 ti is the better choice for sure.
If you don't care about RT or dlss, what's the performance gap? I've heard some people say that the 9070xt beats it in pure rasterization?
-Edit I just watched a benchmark video on YouTube and it seems they're both roughly the same with some games on both side performing a lot better than the other, but pretty much identical everywhere else
The few videos Ive seen where amd is faster has 9070xt at about 3.1ghz vs 5070ti at 2.8ghz. 5070ti can do 3.1ghz no problem if not power limited (base models limited to 300w).
one day AMD fangirls will be honest
need something with 24gb VRAM plus for AI without breaking the bank and isn't the 3090 and isn't 7900xtx.
hodl for 5080 Super, very likely have 24GB VRAM
5080 Super and 5070 Ti pricing will probably be around 700$ gap so I dont think they target the same consumer lol
Asking price for high vram cards in the 2nd hand market now are so grossly inflated. 4090 can go anywhere from 3-4k. Cheapest brand new 5090 is 5.3k.
If they sell the 4080 super with 24gb vram for 2k+, I'll still buy it.
@plentifoo: Yeah, even the 3090 price is gone nuts. I remember seeing used ones get down to like $700 but now they're back up to like $1300.
@Zorololo: I've been offered 2 3090s, $1k each. But prefer to go with newer gen cuda cores.
I mostly use it for stable diffusion.
You do know this is Nvidia yeah? Jensen Ebenezer Scrooge isn't giving people an excuse to not buy a 5090.
There are reports that sales of the 5080 series are doing terribly, and NVIDIA has already cut supply of the 5080 to retailers. The pricing of the 5070 Ti is also ridiculous — I think they’ve completely screwed this up. Just wait, the price drops are going to be fun to watch
The pricing of the 5070 Ti is also ridiculous — I think they’ve completely screwed this up.
Nope, they've not screwed up at all.
Nvidia never wanted to release consumer Blackwell cards at this point in time, the TSMC N4P node is new, expensive, and likely supply limited. They were just forced to because they no longer could produce more 40-series cards (likely TSMC retiring the 4N node). Whilst not exactly identical, the GB202 (RTX 5090) is very, very similar to the GB100 core for datacenters (difference mainly being in the memory interface, and VRAM).
The GB100 goes for upwards of $60,000 each. Each 5090 that Nvidia is producing is basically an opportunity cost of producing a GB100 (or upcoming GB200).
Just wait, the price drops are going to be fun to watch
GeForce is no longer Nvidia's bread and butter (see: https://www.visualcapitalist.com/nvidia-revenue-by-product-l…), their main revenue source is datacenter products (i.e. AI), which is now 78% of their revenue (vs. 17% for GeForce).
Gaming GPUs is basically just a side project at Nvidia now, and it won't be a central part of their corporate strategy going forward. How many GeForce GPUs they sell makes little difference to Nvidia's financial performance, so they most likely will just follow the pattern of pricing high to bite early adopters, then having some small cut with a refresh to keep milking the generation for as long as possible.
The way I personally see it is that Nvidia actually likely see GeForce as just marketing, a way of keeping the Nvidia brand name fresh, and keeping some level of media attention, which may influence / garner corporate / datacenter sales.
GeForce might seem like a side project but I think how many and how much will depend more on wafer pricing and availability. They have a very healthy profit on GeForce, just not nearly as much as data centre because they're at about what consumers can bare while industry has deep investor pockets.
Industry also needs far more varied and rapid software support, as well as hardware reliability, but both architectures are getting enormous software support compared to the pre-RTX years (and thus the pricing involved in early RTX).
But you're right, wafer price and the process R&D involved has been increasing significantly. It will ease a bit with the 3nm options, but 2nm looms after, and then likely a switch to half the reticule limit thereafter, which means MCM goes mainstream.
Nvidia never wanted to release consumer Blackwell cards at this point in time, the TSMC N4P node is new, expensive, and likely supply limited. They were just forced to because they no longer could produce more 40-series cards (likely TSMC retiring the 4N node). Whilst not exactly identical, the GB202 (RTX 5090) is very, very similar to the GB100 core for datacenters (difference mainly being in the memory interface, and VRAM).
This entire statement is nonsense. TSMC would not have retired a thing, NVIDIA just would have moved on from it when updating the Blackwell design, which is now on a custom N4P config (4NP). The same fabs are being used and capacity is already at max, hence the way the 40 series was phased out: they stopped and swapped immediately to 50 series allocations.
GB202 is a derivative of GB100/GB102 in some ways, sure, but it's a big step forward: it's stripped away the 32-bit architecture from Lovelace and early Blackwell, and used AI tooling to redesigning the placement and tracing of the components on the board and chip to better optimise the die area and performance/power consumption as part of a larger re-architecture built to improve AI workload scheduling overtime (the beginnings of neural rendering).
GB100/102 iterates far more on GH100/GH200, and will progress to Blackwell Ultra being offered in its place. None of these have RT cores.
The way I personally see it is that Nvidia actually likely see GeForce as just marketing, a way of keeping the Nvidia brand name fresh, and keeping some level of media attention, which may influence / garner corporate / datacenter sales.
No, they're all people who were once students and gamers, and who also likely see key outcomes for their AI technology advantages over time in gaming, compression, reasonably approximate simulation. Gaming remains a massive market, one they have tried to capture more of previously through an attempt to acquire ARM (though not solely for that reason). NVIDIA may still move into phone GPUs, but are clearly less comfortable with stripping down GPU features and reducing margins to make it happen.
It's not just a fun little fad, and also cloud gaming will present itself as a highly economical solution over time for most potential gamers (and for the environment). It's likely that at the bare minimum, future MMOs, connected open world games, and metaverse engines will have a decent portion of game systems and graphics on the cloud in order to scale things up better.
Entertainment sucks up a huge portion of current internet traffic and spare income, so while digital assistants are going to be great, they're going to be dime-a-dozen (already are) and people are going to want things to do with their free time.
unfortunately it's impossible for nvidia to screw up enough for them to actually give a sh*t… the mind share is extremely strong with consumers as you can see in some of the posts here and how many Si's and OEM's continue to just use nvidia no matter what.
These "cuts" in prices are basically normal pricing movement with increase in supply post launch and the market settling in.. unless AMD keeps it's value gap enough to entice consumers or starts to get their products into OEM/SI builds nvidia will do whatever they want and won't give a crap … especially since Ai makes 100's of times more money for them than gaming does, in their mind they are actually losing money in opportunity cost in making these cards for the gamers so to them they are doing everyone a favor no matter how the consumer is seeing it.
anyone wanna donate this to me? :)
Sure, why not.
9070XT seems like it’s aimed directly at GTA6
Probably 4k 60fps medium settings
should i get this card if i want it use for yolo8?
i'm fairly shallow, its not the prettiest card.
Couldn't be worse than looking in the mirror.
Ouch, it's too early for this savagery.
I dont think this one delivers 100 extra fps than 4070ti
well since the 5070 is a 4090… this card is easily a 5090 at 1/4 of the price…
well since the 5070 is a 4090
Man, that was such a bold faced lie from Jensen, it's crazy.
WOW…Palit is still around? Crazy…
If you didn't know Palit is actually the biggest GPU manufacturer in the world and own other brands like Galax, PNY and Gainward.
Damn, want this but the card is too long for my case (330mm limit)
Sounds like a good excuse to buy a new case LOL
I like the way you think My B550 i gaming is only running 8x PCIE since the latest BIOS update, so I probably need a new motherboard , CPU and RAM too…
So that I can play games a bit faster than I do now..
Hmmm
HODL
Will this run GTA VI?
GTA 6 isn't coming out on PC on launch
GTA6 won't be about killing people, it will be about whether or not you own a house
Kinda tempted to upgrade from my 4070ti but there really doesn’t seem to be enough in it.
Really disappointing generation overall
Same price I paid for my RTX4070TI 12GB in 2022
At least you're getting more VRAM than I did
I eventually sold my 4070TI last year and got a RTX4080 Super 16GB
any reports of melted power connectors on the 5070ti?
HODL