Question How long until RTX 20xx is refreshed?

[RichUK]

2 years? Though suppose 7nm may force a refresh within 2 years.

 

[Furious_Styles]

I thought potentially late this year with 7nm, though that is all the typical BS rumor-mill most likely.

 

[Qwertilot]

Depends on the definition of refresh I'd think. They've currently got the 2080ti out at Titan pricing,and they could theoretically push that down a tier, taking all the other RX ones with it.

If 7nm then a while.

 

[Arkaign]

Compared to 10xx, not too long. But they just came out.

Sometime between fall and next spring is my guess. Navi should force a refresh if nothing else.

The last few releases have been disappointing.

Vega, expensive and less efficient than 1070/1080, and no flagship speed.

RTX, expensive, low on memory, doesn't compare all that hot $/perf to 10xx AIB and even Vega now that it's cheaper.

Vega 7, $$$$$. Not all bad though, I guess 1080ti perf for $699 w/16GB, though if you can find a 1080ti for less, then that's the win.

Ugh.

 

[Dribble]

I don't see Nvidia rushing - they just moved to 12nm, there is no real competition. We'll probably get some gpu compute chips @ 7nm near the end of the year but I can't see the 3 series arriving till well into next year.

 

[Bouowmx]

Feels (gut feeling and not in the basis of any reality) like GeForce 20 series will be a 2-year generation (until 2020 H2, for 7 nm). In the mean time, there will be a HPC replacement for Volta, on 7 nm.

 

[GodisanAtheist]

NV has left themselves some room with cut down parts across the board. Releasing the fully unlocked chips is a freebie. Easy to do by the end of 2019.

Alternatively, a 7nm die shrink, but that's likely a ways off... End of 2020 or further still.

 

[Guru]

Not for at least a year. Even AMD's 7nm GPU's are going to be coming out late Q2 2019, with the first desktop product being the Radeon 7, which is in essence their MI60 workstation graphic card repurposed for desktop.

 

[Mopetar]

Probably depends on AMD. If they just release low-end and mid-range GPUs that don't challenge NVidia at the high end, they don't have much reason to hurry to move down to 7nm. If Navi is a big architectural uplift as well as a 7nm shrink, I think it will cause NVidia to move up any plans they have.

 

[PeterScott]

Probably depends on AMD.

Almost certainly doesn't.

These things have long range schedules with hundreds(potentially thousands) of people in multiple teams working toward an end date. They don't start and stop based on on what the competition is doing.

 

[beginner99]

I expect H12020 or else the release cycle is too short to make a profit of these chips. Also that is when AMD will release Navi and Intel also said they will release first dGPU in 2020. While I expect late 2020 for intel, I don't think NV would risk not having 7nm chips by then.

 

[Mopetar]

Almost certainly doesn't.

These things have long range schedules with hundreds(potentially thousands) of people in multiple teams working toward an end date. They don't start and stop based on on what the competition is doing.

Companies always have plans, but they make adjustments to them based on all manner of factors. For example, NVidia probably delayed production on the 1660 Ti while they cleaned out the remaining 1060 inventory. It's unlikely that they anticipated that in their long range schedule.

The probably have some agreements in place as to how many wafers they'll buy, etc. but they can shift production between different product lines as necessary. Even on a small scale, certain products might have more demand requiring such adjustments.

Another thread just indicated that they made a pretty big adjustment to their earning expectations. Clearly whatever plans they made have been thrown off by a fair bit, which is going to necessitate changing them. Just because you set a plan and changing it (for the reasons that you point out) might be difficult, doesn't mean that a company must stick with it, especially if the conditions under which that plan was formed don't seem to hold.

 

[Headfoot]

I think we will see a 20xx refresh fairly soon actually, end of q4 2019 or q1 2020. My reasoning is that their original design target was 7nm, not 12. They backported a design to 12nm to get viable yields, cost, time to market etc. But no way they just stopped working on the 7nm design. They're certainly using that as extra time to refine and enhance the existing 7nm design.

IMO the wildcard is whether they hold back the consumer version of it to recoup on 20xx and only release the 7nm refreshed version for professional lines for 6+ month. This is the lever that might be pulled by competition, but the design work is certainly independent

 

[AtenRa]

With current NV news (lower Turing sales) and AMD ramping up 7nm production in 2019 (CPUs + GPUs) , I dont see NV releasing a consumer 7nm product this year.

 

[PeterScott]

I think we will see a 20xx refresh fairly soon actually, end of q4 2019 or q1 2020. My reasoning is that their original design target was 7nm, not 12. They backported a design to 12nm to get viable yields, cost, time to market etc. But no way they just stopped working on the 7nm design. They're certainly using that as extra time to refine and enhance the existing 7nm design.

IMO the wildcard is whether they hold back the consumer version of it to recoup on 20xx and only release the 7nm refreshed version for professional lines for 6+ month. This is the lever that might be pulled by competition, but the design work is certainly independent

Nah, they are still rolling out 12nm, with a new die for GTX 1660, and then possibly another new die below that. 2020 sure, but RTX started shipping in 2018, so 2020 is potentially 2 years later, which is just a normal release cycle.