4070TI reviews thread

[BFG10K]

NVIDIA France Accidentally Reveals GeForce RTX 4070 Ti Specs

With less than a week to go until the official launch of the GeForce RTX 4070 Ti, NVIDIA France has gone and spoiled things by revealing the official specs of the upcoming GPU. The French division of NVIDIA appears to have posted the full product page, but it has since then been pulled. That...

www.techpowerup.com

First official performance leak. A perfect opportunity to highlight the fraud pushed by nVidia with DLSS.

Performance looks fantastic, right? Look again, at the smallprint - 4070 is using frame generation.

I said exactly this would happen as soon as DLSS 3.0 was announced. 2560x1440 is a lie given it's actually ~1080p due to legacy DLSS. Since nVidia got away with it, the next step is frame interpolation lies. It's "faster", yo!

Adjusting for frame generation (i.e. halve the 4070's bars), it's actually barely faster than the 3080 in MFS, and slower in Warhammer 40K. All this from a card that will likely cost more.

Customers not seeing the smallprint and/or not understanding DLSS are having systematic fraud perpetuated on them.


Reviews

MSI GeForce RTX 4070 Ti Gaming X Review

MSI's GeForce RTX 4070 Ti Gaming X comes with a large triple-fan, triple-slot thermal solution that runs at whisper-quiet noise levels with low temperatures. In terms of performance the new card is able to match last generation's flagship, the RTX 3090 Ti, and offers new features like DLSS 3...

www.techpowerup.com

 

[moonbogg]

50% off

Yep, 70 class should be in the $400's. There, I said it. And if people say they can't make one for that price, then at that point the product no longer makes sense for what it is. Imagine Sony making a playstation that costs $1500. Is there anything they could say to justify it? Nope. Not a thing.

 

[SteveGrabowski]

Is there a better $800 card?

Is this not a huge jump in price from typical 70 series releases? And it's also kind of weak by what the 70 series has historically been vs the previous generation. Might as well argue that we should have exponentially increasing gpu prices that track with performance improvements.

 

[Stuka87]

Is this not a huge jump in price from typical 70 series releases? And it's also kind of weak by what the 70 series has historically been vs the previous generation. Might as well argue that we should have exponentially increasing gpu prices that track with performance improvements.

It launched at a higher price that the 3080 launched at. 3080 was $699, the 4070 is $799.

 

[Topweasel]

Is there a better $800 card?

Maybe not. Though a 6950 a week or two ago would have been pretty close. But its not just about how it compares as the lowest specced of current gen releases. Lovelace is fast enough that just being the lowest price of available newer cards still makes it faster than most last gen products (minus the 6950, 3090, 3090ti). But its also what it does to the market, its a significant shift in pricing for this market of card and it wasn't 2 months ago Nvidia intended to sell it for $100 more and pretend it was a match (outside memory) for a card 30% faster. Nvidia and AMD (because they don't see themselves as value brand) are fine with trying to push through with mining premiums, even though the market has fallen out.

People purchasing it just because they have $800 and this is the fastest $800 card are just making the situation worse. AMD for example is going to have to do some significant price cuts considering how much even Intel is eating up the market share they had. This isn't a setting where could sell the dies for other products. They have to move them. If AMD does, then that would put pressure on Nvidia to do as well. If we keep the market at all time lows. Waiting for price cuts. They will come.

 

[Heartbreaker]

Is this not a huge jump in price from typical 70 series releases? And it's also kind of weak by what the 70 series has historically been vs the previous generation. Might as well argue that we should have exponentially increasing gpu prices that track with performance improvements.

That's merely semantics. Naming doesn't matter.

It's still the best $800 card you can buy. It's perf/$ is in line with other cards of this generation.

I get that some want it to $400, but I'd like Halle Berry to be my sugar mamma, with a similar connection to reality.

 

[NIGELG]

That's merely semantics. Naming doesn't matter.

It's still the best $800 card you can buy. It's perf/$ is in line with other cards of this generation.

I get that some want it to $400, but I'd like Halle Berry to be my sugar mamma, with a similar connection to reality.

Stop Defending Nvidia AND AMD's ridiculous pricing.

Maybe Intel can throw a spanner in the works. The true midrange gamer is now out in the cold with these prices.

 

[linkgoron]

That's merely semantics. Naming doesn't matter.

It's still the best $800 card you can buy. It's perf/$ is in line with other cards of this generation.

I get that some want it to $400, but I'd like Halle Berry to be my sugar mamma, with a similar connection to reality.

LOL this is such a bad take. Using this logic we should've had GPUs and CPUs at 10x the price they currently are. It's also completely out of sync with how the PC market has been behaving since forever.

You used to get more for less when technology moved forward, but now suddenly we're fine with the same for the same?

 

[Heartbreaker]

Stop Defending Nvidia AND AMD's ridiculous pricing.

We don't have a clue about the actual cost to produce, just guesses. If the prices were that ridiculous, you would think AMD with it's tanking market share would try to win some back by lowering it's prices.

R&D Intensive companies like AMD and NVidia, need about 40% gross profit margins (you can check historical profit margins, anytime either company went below 40% the company lost money), just to break even as a company, plus there will be some additional small margins needed for the AIBs, and then some small margin for the retail chain...

Anyone who thinks they can just cut prices in half and maintain a healthy business, is living in a complete dreamworld.

 

[Heartbreaker]

Funny 4070 Ti message from MSI. It's price is "not as bad as a 4080"!

MSI says GeForce RTX 4070 Ti cost is 'not as bad as RTX 4080' - VideoCardz.com

MSI admits GeForce RTX 4080 pricing is bad Even MSI is not happy with the RTX 4080 pricing. MSI Gaming USA happily reports that RTX 4070 Ti is now available at their official store. Interestingly, RTX 4070 Ti is still available at MSRP 3 days after launch, even at this store specifically which...

videocardz.com

 

[coercitiv]

R&D Intensive companies like AMD and NVidia, need about 40% gross profit margins (you can check historical profit margins, anytime either company went below 40% the company lost money), just to break even as a company, plus there will be some additional small margins needed for the AIBs, and then some small margin for the retail chain...

Anyone who thinks they can just cut prices in half and maintain a healthy business, is living in a complete dreamworld.

Here we go again with the production cost fallacy. You're the one living in a complete dreamworld where total costs for GPUs have increased by 100% in 5 years.

Meanwhile... surprise surprise, laptops have remained just as affordable (unless they feature a strong dGPU). I guess laptop makers are living in a dreamworld too.

 

[jpiniero]

Here we go again with the production cost fallacy. You're the one living in a complete dreamworld where total costs for GPUs have increased by 100% in 5 years.

Meanwhile... surprise surprise, laptops have remained just as affordable (unless they feature a strong dGPU). I guess laptop makers are living in a dreamworld too.

10 nm blows but Intel is still getting relatively reduced costs via depreciation. AMD is going to have a hard time competing to the point where (as AT put it) Cezanne is back for round 3.

Even then Intel is hiking prices 15%. $599->$799 is 33% but you are talking about wafer prices that yes are 100% or more higher compared to SS8.

 

[Heartbreaker]

Here we go again with the production cost fallacy. You're the one living in a complete dreamworld where total costs for GPUs have increased by 100% in 5 years.

Wafer costs have increased substantially, and GPU's are typically much larger than CPUs, so will feel the pain of this more.

5 years ago: NVidia Gross Margins 59%, most recent Quarter 58%.
5 years ago: AMD Gross Margins 42%, most recent Quarter 46%.

This does not reflect companies enjoying a massive margin increase.

 

[NIGELG]

Well then PC gaming will be a niche thing.

 

[Heartbreaker]

Well then PC gaming will be a niche thing.

You can PC game on an IGP. If I could get an AMD Phoenix APU, I'd probably just game on that. It's better than my current discrete card.

Gaming on higher END GPUs is a niche thing, and has always been. If you look at the Steam HW survey, it's completely dominated by 50 and 60 segment cards.

 

[Ranulf]

Well, maybe your point will hold when they finally release the x50 and x60 level cards this year...maybe by summer. Which is not the norm. Its been 3 months now and I don't see or hear either company talking about the mid range at CES.

If those new x60 or x60ti (which will be released first by Nvidia) cards or the 7600XT etc. are $500? Will they even be as fast as a 6700XT, 6800XT or 6900XT? At this rate the 4060ti will be barely faster than a 3060ti or 3070. Stagnation central.

 

[gdansk]

The RTX 3080 and 6800XT MSRPs were "too good" seems to be AMD's and Nvidia's stance. But there is no doubt that tiny AD104, with great yields, are cheaper and easier to produce than GA102 was. Improved supply and reduced demand should mean the prices drop but here we are. Maybe it'll sort itself out in a few months.

 

[Mopetar]

There's still too much left over inventory in the channel from the previous generation for Nvidia to want to release lower tier cards. Far better to devote wafers to AD102 and sell high-margin 4090 cards and start stock-piling 4080 Ti cards for later.

 

[SteveGrabowski]

That's merely semantics. Naming doesn't matter.

It's still the best $800 card you can buy. It's perf/$ is in line with other cards of this generation.

I get that some want it to $400, but I'd like Halle Berry to be my sugar mamma, with a similar connection to reality.

So disregard the fact that this generation is wildly expensive compared to every other generation we have seen? At these prices better to just wait since no one is buying gpus anyways.

 

[Heartbreaker]

So disregard the fact that this generation is wildly expensive compared to every other generation we have seen? At these prices better to just wait since no one is buying gpus anyways.

It's still an improvement in perf/$ which is is more important than semantic arguments. Perf/$ changes are similar to what we saw with Turing.

IMO there will be much more legitimate complaints at lower performance levels where their will be much more negligible improvements.

 

[SteveGrabowski]

It's still an improvement in perf/$ which is is more important than semantic arguments. Perf/$ changes are similar to what we saw with Turing.

IMO there will be much more legitimate complaints at lower performance levels where their will be much more negligible improvements.

And Turing got similarly tepid reception from gamers who skipped it.

 

[SteveGrabowski]

LOL MSI crapping on the 4000 series

https://twitter.com/x/status/1611500515682291712

 

[Heartbreaker]

LOL MSI crapping on the 4000 series

Yeah, already posted that up the page. It's specifically taking a shot at the 4080, and daming the 4070 ti with faint praise (not as bad as the 4080).

And Turing got similarly tepid reception from gamers who skipped it.

That's fine. Did I say you should rush out and buy one?

I just don't think the 4070 Ti is a reason to melt down. It's the best perf/$ of Ada.

 

[tajoh111]

Here we go again with the production cost fallacy. You're the one living in a complete dreamworld where total costs for GPUs have increased by 100% in 5 years.

Meanwhile... surprise surprise, laptops have remained just as affordable (unless they feature a strong dGPU). I guess laptop makers are living in a dreamworld too.

NVIDIA Cost of Goods Sold 2010-2024 | NVDA

NVIDIA annual/quarterly cost of goods sold history and growth rate from 2010 to 2024. Cost of goods sold can be defined as the difference between beginning and ending inventories for tangible products resulting in an expense that reflects production and sales costs. <ul...

www.macrotrends.net

NVIDIA Research and Development Expenses 2010-2024 | NVDA

NVIDIA annual/quarterly research and development expenses history and growth rate from 2010 to 2024. Research and development expenses can be defined as an expense arising from studies and product development processes. <ul style='margin-top:10px;'> <li>NVIDIA research and...

www.macrotrends.net

R and D and cost of production have doubled every 2 years. So while they could keep prices the same, profits and margins would get smaller. Product releases would have to become slower and performance increases smaller. Cost have to be passed onto the consumer or spending has to be cut to be profitable.

Pragmatically speaking, considering Nvidia has very little power to fight TSMC. Thus the only way Nvidia can save money is on R and D spending.

To do this, Nvidia needs to get rid of more of their North American Engineers and outsource more to China. There the Engineers makes 20 percent of what they make in North America. AMD did this during the save the CPU side at the expense of GPU division which is why it could accomplish so much with a little budget. The engineers being produced out of China are top notch too. It took a while but AMD Shanghai is making some great products now.

One of the reasons why CPU have not had a dramatic increase in cost is core counts have not increased as of late and CPU have shrunken in size. Nvidia or AMD could keep the core counts of their GPU static and use only architecture and increases in the node to increase performance but we would quickly reach power limits and get largely small increases in performance.

I think cards are overpriced but I think 50 percent of it is too much Amphere inventory and the other 50 is the increased cost of production.

I think realistic pricing if AMPHERE stock was not a factor would have been $699 for the RTX 4070 ti, $899 for the RTX 4080 and the same pricing for the RTX 4090 but perhaps a little less cut down.

Not great pricing but fair when you include the cost of production increase.

 

[gdansk]

Cost of goods sold is increasing slower than revenue. Says it all really.

The only Ada that is more expensive to produce than its predecessor is the RTX 4090. The rest are tiny and therefore cheap even with the highest estimates for 5nm wafer price. So this trend will continue.

 

[SteveGrabowski]

Yeah, already posted that up the page. It's specifically taking a shot at the 4080, and daming the 4070 ti with faint praise (not as bad as the 4080).




That's fine. Did I say you should rush out and buy one?

I just don't think the 4070 Ti is a reason to melt down. It's the best perf/$ of Ada.

Why is not liking the 4000 series considered melting down lol?