- Nvidia released two budget offerings priced at $109 and $139.
- In response, AMD slashed the prices of its RX460 and RX470 GPUs by $10 each.
- But this price cut may not be enough for AMD to curb Nvidia's incursion into its territory.
Anyone actively following the GPU trends would know by now that Nvidia introduced two new Pascal-based budget graphic cards last week, priced at just $109 and $139, in a bid to nab AMD's (NASDAQ:AMD) market share. The cards aren't even out yet and AMD, in what seems to be an overreaction, has already slashed the prices of its RX460 and RX470 cards by $10 each. A few might expect this abrupt discounting of GPUs to dampen Nvidia's blow on AMD, and save the latter from a market share erosion, but there's reason to believe that it won't. Let's take a closer look at things to have a better understanding of it all.
Let me start by saying that Nvidia GPUs have historically retailed at a premium compared to AMD offerings as the former's offerings have generally been more performance oriented. AMD GPUs may have fared decently in terms of comparative performance-per-dollar or performance-per-watt metrics, but they tended to generally lag similarly spec'd offerings from their larger rival in terms of raw performance. So this market dynamic pretty much created a distinction between the two companies; Nvidia dominated the high-end GPU segment whilst AMD enjoyed the low-end of the market.
But all this will seemingly change with Nvidia's decision to go downmarket, aggressively enter AMD's strong segment and release products that would go head-to-head with AMD cards in terms of performance-per-dollar and performance-per-watt metrics. AMD tried to counter this insurgency by slashing the prices of its RX460 and RX470 GPUs by $10 each. On second thoughts, the word "slashing" might be an overstatement in itself since its only a $10 reduction.
The $10 reduction on price tags of $110 and $180 may appear to be a significant discount to a few in terms of absolute percentages; for starters, the discount equates to about 9% and 5% discounts on RX460 and RX470 respectively. But this isn't a significant reduction by any means. To put things in perspective, there's just a $9 difference between AMD's RX460 and Nvidia's new GTX 1050. Mark Walton at Arstechnica believes that that Nvidia's 1050TI costing $139 would be slightly faster than the 4GB variant of AMD's RX460 that costs $129.
Now why would anyone want to compromise on performance for just $9 or $10? As I said earlier, AMD's discounts may look appealing on a percentage basis, but it's a really marginal difference compared to Nvidia's offerings. The trade-off between price and performance works in Nvidia's favor, as very few would actually base their purchasing decision on the $10 differential. Hence, I'm convinced that AMD won't be able to control Nvidia's onslaught in the low-end GPU segment with its meager price cut.
So what can AMD do?
An option for AMD would be to take the fight back to Nvidia’s camp. The former’s offerings have generally been discounted compared to Nvidia’s and released with an extensive focus on performance-per-dollar metric. But that could change if the chipmaker went upmarket.
AMD could simply introduce one or two overclocked SKUs based on the Polaris architecture, that aren’t geared to keep the overall electricity draw in check, but rather built to deliver significantly higher performance than its current flagships. Gamers in the high-end graphics segment generally care less about power-draw and more about raw performance so maybe AMD can take advantage of this market dynamic.
This way, the incremental R&D spend of developing an entirely new line of GPUs won’t be involved. I believe the costs would actually be minimal if the chipmaker heads this way, as AMD would merely be developing an overclocked version of its existing GPUs, housing them with electronic and mechanical components well equipped to handle higher temperatures and power draw.
Perfecting the Vega release could also help
The suggestion mentioned above discusses tweaking an aspect of AMD GPUs based on Polaris architecture. What we didn’t discuss yet is AMD’s upcoming Vega architecture. Both AMD and Nvidia are scheduled to release their Vega and Volta architecture-based GPUs next year, respectively. If AMD doesn’t fix the aforementioned pricing problem and repeats the same mistakes made with the launch of Polaris GPUs, Nvidia is most likely going to intensify the battle in the low-range with its next-gen GPUs.
For starters, there was power-surge related issue that reduced the appeal factor of its cards. AMD’s forte is its appealing performance-per-watt metrics, but when its clientele and market testers saw that the GPUs were drawing more power than specified, they were disappointed. So AMD will have to address this technical snag, which hopefully should end with Polaris architecture, and not continue into Vega as well.
AMD needs to also compete with Nvidia in terms of raw performance. It must consider launching a few GPU SKUs that aren’t limited in power due to electricity draw considerations. Top end gamers rarely worry if their GPUs are drawing 20-30W extra, but they do care if their 4k games are playable at decent fps or not. Doing this would open up the high-end GPU segment for AMD.
And lastly, AMD needs to release low, mid and high-end Vega chips in a timely manner, preferably in a quick succession. Delaying the launch of any segment would simply hurt its sales and hand over the market to Nvidia. AMD must be on the offensive, releasing products proactively, and any sort of delays in its next year’s GPU offerings would simply extend Nvidia’s lead over its small rival.
Putting it all together
Nvidia’s decision to aggressively go downmarket would certainly hurt AMD for the remainder of this year, and there is very little that AMD can do right now to curb the incursion, except for slashing its prices or releasing a few overclocked GPUs. So investors need to keep an eye on AMD’s Vega release. The chipmaker can take the fight back to Nvidia if it plays cards right. Don’t rule out AMD just yet.