Cheapest price I've seen for the 4070 Ti so far, but there's still a few hours left in the day.
8 in stock.
1% surcharge for Card & PayPal payments.
I missed it previously, but this was posted on the 2nd so it's showing up as a duplicate even though it's expired. Reposting because it's un-expired and people might want to know.
Still too expensive
There had been many $1299 deals already https://www.ozbargain.com.au/product/nvidia-geforce-rtx-4070…, I don't know why this price is attractive, especially when you add the surcharge which brings it to $1311.99
While I agree this isn't an amazing price for a 4070 Ti, this is the Gaming X Trio which has been this price on Shopping Express before, but usually this model is around $100 more than other cheaper 4070 Ti's.
$1430 is the next cheapest price I can find this specific model: https://www.ccpu.com.au/show_prod.php?class_id=video-nvidia&…
the 40 series really don't matter what AIC models they are, they all perform pretty much the same within +/- 5%, reason being https://www.ozbargain.com.au/comment/13419741/redir
I can see there are a few $1290 cards if anyone wants to grab, but again, they are normal price rn:
https://www.staticice.com.au/cgi-bin/search.cgi?q=Rtx+4070+t…
While I agree they all perform nearly identically, the fact is PNY and Inno are not well known brands here and many people are still willing to pay more for a known brand like this.
the 40 series really don't matter what AIC models they are, they all perform pretty much the same within +/- 5%, reason being
I'm not disputing that most AIB models perform the same. All I'm saying is that this model is usually $100 more. Some people want to buy a particular brand for the look/aesthetic, warranty process, brand reliability etc. Could be any reason, regardless, this card is usually $100+ more than PNY or INNO3D cards. Mind you I have a GALAX card in my system right now and haven't noticed any difference from when I had an EVGA card (RiP to the GOAT NVIDIA brand).
@KARMAAA: I’m being lazy by not searching but I’m curious - is there a chart or something outlining the warranty process for various brands in Australia? Eg: Is it to be sent overseas, do they provide an advance exchange, etc?
@2024: Not really, many brands don't tell you these sorts of things until you have to go through the process. They do it on purpose imo. But 9/10 times just take it back to the retailer and they will do the RMA/warranty process for you and you won't have an issue.
What I can say is that EVGA had the best warranty process and also policy, they would allow for global warranty (as-in, if you moved overseas with your PC they would honor your warranty anywhere in the world) and they also allowed for second-hand buyers to enjoy warranty. They also pretty much had a no questions asked warranty policy, meaning you could replace paste, do overclocking (as long as you didn't do like some mod like solder resistors to your board for extended power limit), flash a different BIOS (at least for cards with two BIOS') and have your card covered.
Most brands won't even cover you if you re-paste your card which is kind of stupid (but I understand why it's because you could potentially ruin some component in the process, yes there are customers out there like this I've seen it at PC shops all the time who somehow ruin something which is pretty simple), or you have to argue with them about it over and over and bring up the ACCC or take them to small claims tribunal to get your card's warranty. It's just the way things are. I miss EVGA dearly because they were simply the best when it came to this stuff but they had to move on from NVIDIA's greed. But yeah pretty much it's a lucky dip/draw of the straws as to whether your situation will be covered under warranty. But most places make you ship to somewhere in AU and cover the cost for your shipping.
I heard this one is pretty quiet
Curious, worth upgrading from a 3070 to this?
Nope
No easy one word answer tbh, depends on what you play and at what resolution. Do you want to wait half a year+ for possible $100/$200 off or not. It's more performance than a 3070 at a price that makes some people uncomfortable, but in Australia it's by far the more attractive 4000 gpu price wise as 3000 is now out of production and buying used comes with its own can of worms.
If your answer is yes I want more performance at 1440p in my AAA titles at highest settings and you want it now at a slightly higher than normal price (personally I think this same card at $999 would be a tough thing to say no to) then go for it. The triple X by MSI is a pretty good cooler design and my previous MSI cards had next to no coil whine either.
If you are still getting at least 60fps for the games you play at the screen resolution you use then the answer is likely no. If getting less than 60fps average then the answer starts to become yes. Of course everyone's base level of acceptable performance differs, but that's a guide.
Here's a slightly cheaper alternative:
[eBay Plus] Gigabyte GeForce RTX 4070 Ti 12GB GDDR6X Gaming OC $1,284.84 Shipped @ Futu_online via eBay
Had considered posting as a deal but discarded it.
Pro-tip: Afterpay day is coming up this month. If you aren't in a hurry, wait for AFP and see if there are better deals then.
Afterpay Day Sale starts on Thursday, March 16th 2023 and ends on Sunday 19th March 2023.
thanks for the heads up . Let's hope this year they actually have some decent card deals!
You can get solid full 4070 Ti pre-builts for $500 more than this right now, which tells you this price is about to drop.
that only tells you that the wholesale price for graphics cards are much cheaper, and both Nvidia and the retailers are ripping us off with huge margin…..
No, it's telling you that NVIDIA set the MSRP that way to sell 30 series cards, and are now promising rebates in the not-too-distant future for retailers to shift the pricing structure. This is probably 4 months behind the schedule they set for this originally.
I don't think they ever wanted to launch the 4080 at the price it was, they've just had to help a bunch of board partners shift inventory that they overbought purely to hand over to crypto miners, who for obvious reasons disappeared when the bubble burst. Seems such vendors were expecting things to hold until ETH mining was no longer viable, except they didn't bother to check how hard it was to mine well before that.
It's a move that NVIDIA can never make again now, but their little side benefit is they can test the waters of a $799 RTX 4080, which helps cover the increased wafer costs TSMC threw at them at the last minute, and the opportunity for more margins in the future. Sure, this would have helped NVIDIA out a bit as manufacturing processes ramped and went through early teething issues, but no one's going to let such a heft early adopter tax fly in the future.
…AMD also launched at much higher price so what's the excuse there, they aren't drowning in over supply.
The higher price wasn't an accident by launching at a higher price they've still driven the overall prices up and next gen when they launch at a xx80 series at $1599 everyone will cream themselves cos its so much better than last gens price when in actuality its now locked in at $400+ more expensive than the 3080 was.
@botchie: Not if the suckers keep buying them. Thankfully it does look like sucker fever has slightly diminished.
Maybe rate rises do stop the poor from driving prices up. At least with GPU's
@JimmyNegger: its just mining bust, people could have bought a card for $1500 in the past and it would pay itself off in 4 months, was crazy, and ofc prices went up to the moon, now not so and lets hope it never comes back again
@Osiris: AMD were baited into thinking that NVIDIA were going to deliver a chunk of their performance increases through massively increased power draw due to the numbers coming out of their initial testing, so AMD took advantage of TSMC's process to do the same, which would have covered for the performance penalties in their first gen MCM consumer GPU design.
Then NVIDIA was revealed to be reigning in power limits, and in the end even turned out to change their definition of TDP to simply indicate peak power consumption. Suddenly AMD are also winding back power consumption, only their performance numbers turned out to be based on something like 500+ W power draw.
So AMD were left with a design that didn't really stretch much further gen on gen - though it does perform well in some less latency and clock-sensitive scenarios - but was inherently more expensive. Their pricing for Navi 31 reflects that for the design and manufacturing costs, they expected it to more thoroughly compete with the 4090, and have had to cut into their expected margins fairly substantially.
Meanwhile NVIDIA are whacking them over the head with a board design that is mostly a traditional xx70 series card, just with double the usual VRAM and larger-than-typical and densely packed chip that has plenty of cache. In essence, NVIDIA made use of the same techniques that saw RDNA2 be competitive (and later, affordable), but also got to make 2 process leaps and an AI-optimised trace design pass at the same time.
RDNA 4 should see AMD fix a lot of mistakes and give them a chance to improve RT and ML acceleration, but if NVIDIA have to go to MCM next gen, they've got some designs ready to go that AMD will again struggle to compete with unless they get creative. Also while this is all happening, Intel are coming at a rate of knots in the midrange after making a couple of design errors in their first attempt.
The TLDR is that AMD's GPU team just isn't pushing the boundaries as hard as their CPU team is right now, though the latter is going to hit some rough water soon due to complex hybrid chiplet/SoC designs becoming the norm, somewhere that AMD seems to be behind the entire pack unless they've got some surprises with Zen 5-based desktop designs.
@jasswolf: You wrote a lot of text there but your entire post doesn't discuss anything related to supply and prices and it almost sounds like you are trying to absolve AMD this just rather than accepting they are just like Nvidia. Neither of them are our friends, its a duopoly of 2 corporations whose purpose is to extract as much profit as possible and both of them have opted for higher prices it cannot be argued otherwise.
@Osiris: It absolutely can be explained away by card supply, which is only now finishing up draining GA104. $2200 4080 pre-builts and $1788 4070 Ti pre-builts also highlight that prices are now fully correcting for NVIDIA cards in a way almost exactly like AMD priced the RX6000 series at launch, once you adjust for current TSMC 4nm wafer costs.
Next gen prices will be beholden to TSMC 3nm or Samsung 3GAP wafer costs, of course, but for the former we're talking 25% cost increase for say 65% increased density (subject to SRAM requirements, which barely change in density with this node), so there's plenty of scope to push another monolithic design even at the top end.
So let's stick with monolithic and take a worst-case scenario: the AD102 successor is made on TSMC 3nm, with wafer costs still at $20,000 a full year after manufacturing ramp, with NVIDIA having planned for this and capping out at around a 480 sqmm die, which without making any substantial architectural changes would result in a 190 SM chip.
I personally think TSMC wafer prices will drop as the process matures and we should see something close to a 600 sqmm die again, which leaves a lot of architectural scope to build out while still pushing key subunit quantities. Clocks won't really go much further - if at all - and the focus should be on pushing RT and ML upscaling performance relative to the rest of the architecture, particularly raytracing.
it almost sounds like you are trying to absolve AMD this just rather than accepting they are just like Nvidia.
Absolutely not, RDNA3 is not good, but it is a first generation chiplet GPU design.
Neither of them are our friends, its a duopoly of 2 corporations
Three, if you now count Intel, six if you now consider Apple, ARM and Qualcomm as competitors in the low and lower midrange, which they increasingly will be.
whose purpose is to extract as much profit as possible and both of them have opted for higher prices it cannot be argued otherwise.
AMD's pricing structure has so far been the same in terms of the top Navi 31 SKU at launch vs Navi 21: $999. The packaging costs of the chiplet design meant the next SKU was also similarly expensive, that's all. These are halo cards, but the nature of chiplet design makes the base cost of any cutdown version more expensive than everyone is used to.
It'll be the same thing with Navi 32, subject to the number of MCDs required per SKU.
This is one of the most well thought out posts I have read, not just on OzB but anywhere. You either have a good mind or you know a bit about the industry.
Will update my 6800 when new cards is 2-3 time’s performance and under $1000. I prefer not to pay too much to greedy GPU manufacturers. It is okay they keep the price of their GPU, I will keep my money.
no need to upgrade 6800 xt at all - unless you want 4k 144 hz in which case you get 4090
Shouldn't these be $7-800 by now? FML!
They should be $1000 soonish, but the real question I think you're angling for is what will the specs of the RTX 4070 be. Current rumours say it's a big drop, but then we should be expecting the cheapest xx70 MSRP that we've seen for years.
I think the specs will be pretty close to the 4070 Ti.
Ima hold longer lol until i have a real need for a new gpu. now i only want one.
AMD and NVIDIA, Crypto boom and chip shortages is over. No more lock down and people are back in the office, so demand for GPU has dropped significantly, and people ain't stupid enough to pay for this ridiculous high price.
Isnt the 4070 the scam one?
Smaller Memory Bus 192 bit
504Gb/s
Definitely a cut back card
You can find "Gigabyte GeForce RTX 4070 Ti 12GB GDDR6X Gaming OC" for $1284.84 with eBay Plus 17% off: https://www.ebay.com.au/itm/334691056578