Just weeks after unveiling its 40-series cards powered by the all-new Ada Lovelace architecture, Nvidia said Friday it was pulling the plug on the RTX 4080 12GB before the card had even hit store shelves.
“The RTX 4080 12GB is a fantastic graphics card, but it’s not named right. Having two GPUs with the 4080 designation is confusing. So, we’re pressing the ‘unlaunch’ button,” a brief news release read.
RTX 4080
Nvidia still plans to move forward with the more powerful 16GB version of the card, which it says will go on sale November 16.
Announced at Nvidia’s GTC event this fall, the RTX 4080 was to be available in two versions: a 12GB model slated to retail for $899 and a 16GB model at $1,199. This isn’t all that unusual for Nvidia, which sold 10GB and 12GB versions of the 3080 in addition to the higher-spec 3080 TI.
However, a closer inspection revealed that the RTX 4080 cards were separated by more than a few hundreds dollars and 4GB of extra vRAM. It turned out the 12GB 4080 used a different GPU die — the AD104 — with 7,680 CUDA cores. That’s nearly 2,000 fewer CUDA cores than the 16GB model at 9,728. What’s more, the 12GB 4080’s memory bus was also 25 percent smaller.
- Ethereum Merge signals end of GPU shortage, but not necessarily high prices
- Nvidia unveils RTX 4090 – but it’s the 4080 to watch out for
- PC shipments fall at fastest rate ever as businesses slam wallets shut
- The new GPU world order is beginning to take shape
A review of past Nvidia releases show the US giant has typically reserved its 104 dies for its 70 cards, like the RTX 2070 and 3070. This suggests that we may see the 12GB card return in the form of a RTX 4070 at some point in the future. One might argue the 12GB 4080 should have been a 4070 in the first place.
It’s unclear how this might affect Nvidia’s board partners so close to the November launch date of the new series. Regardless, the decision means Nvidia will head into the holiday shopping season with its least expensive 40-series card retailing for $1,199.
However, this may be intentional. By keeping the price of the 40-series high and holding off on launching lower cost cards, Nvidia may be trying to avoid prematurely diluting the value of its piles of leftover 30-series stock, which it really needs to shift.
It may be a combination of bad branding and inventory issues, Moor Insights and Strategy analyst Anshel Sag told The Register. “Inventory isn’t moving quickly enough, especially 3090s.”
RTX 30-series pricing has steadily crept downward over the past few weeks, following the collapse of the crypto market and Ethereum’s transition to proof of stake. Ampere-based RTX 3090s are retailing for less than $1,000 at many online retailers and 3080s can be had for at or close to MSRP.
That ought to help Nvidia sell that Ampere stock sitting in warehouses.
At the same time, PC demand continues to erode. Just this week, Gartner reported PC shipments slid nearly 20 percent from this time last year, the steepest decline since the mid-1990s. ®
Just weeks after unveiling its 40-series cards powered by the all-new Ada Lovelace architecture, Nvidia said Friday it was pulling the plug on the RTX 4080 12GB before the card had even hit store shelves.
“The RTX 4080 12GB is a fantastic graphics card, but it’s not named right. Having two GPUs with the 4080 designation is confusing. So, we’re pressing the ‘unlaunch’ button,” a brief news release read.
RTX 4080
Nvidia still plans to move forward with the more powerful 16GB version of the card, which it says will go on sale November 16.
Announced at Nvidia’s GTC event this fall, the RTX 4080 was to be available in two versions: a 12GB model slated to retail for $899 and a 16GB model at $1,199. This isn’t all that unusual for Nvidia, which sold 10GB and 12GB versions of the 3080 in addition to the higher-spec 3080 TI.
However, a closer inspection revealed that the RTX 4080 cards were separated by more than a few hundreds dollars and 4GB of extra vRAM. It turned out the 12GB 4080 used a different GPU die — the AD104 — with 7,680 CUDA cores. That’s nearly 2,000 fewer CUDA cores than the 16GB model at 9,728. What’s more, the 12GB 4080’s memory bus was also 25 percent smaller.
- Ethereum Merge signals end of GPU shortage, but not necessarily high prices
- Nvidia unveils RTX 4090 – but it’s the 4080 to watch out for
- PC shipments fall at fastest rate ever as businesses slam wallets shut
- The new GPU world order is beginning to take shape
A review of past Nvidia releases show the US giant has typically reserved its 104 dies for its 70 cards, like the RTX 2070 and 3070. This suggests that we may see the 12GB card return in the form of a RTX 4070 at some point in the future. One might argue the 12GB 4080 should have been a 4070 in the first place.
It’s unclear how this might affect Nvidia’s board partners so close to the November launch date of the new series. Regardless, the decision means Nvidia will head into the holiday shopping season with its least expensive 40-series card retailing for $1,199.
However, this may be intentional. By keeping the price of the 40-series high and holding off on launching lower cost cards, Nvidia may be trying to avoid prematurely diluting the value of its piles of leftover 30-series stock, which it really needs to shift.
It may be a combination of bad branding and inventory issues, Moor Insights and Strategy analyst Anshel Sag told The Register. “Inventory isn’t moving quickly enough, especially 3090s.”
RTX 30-series pricing has steadily crept downward over the past few weeks, following the collapse of the crypto market and Ethereum’s transition to proof of stake. Ampere-based RTX 3090s are retailing for less than $1,000 at many online retailers and 3080s can be had for at or close to MSRP.
That ought to help Nvidia sell that Ampere stock sitting in warehouses.
At the same time, PC demand continues to erode. Just this week, Gartner reported PC shipments slid nearly 20 percent from this time last year, the steepest decline since the mid-1990s. ®
