9 September 2024

AMD's New Reality...


Last time, I pontificated on the current situation surrounding two of the gaming industry greats. Today, I'm going to address very recent comments on their current strategies relating to the consumer (client) graphics card market...


It may not be what you think it is...


The Setup...


Recently, Jack Hyunh spoke with Tom's Hardware and outlined some of AMD's broad-strokes reasoning relating to the consumer GPU market.

The issue, here, is that Jack specifically came out with a couple of choice quotes relating to strategy of client (i.e. "us") devices, going forward:
"Of course, we have to because that’s performance-per-dollar. Even Microsoft said Chat GPT4 runs the fastest on MI300. Here's the thing: In the server space, when we have absolute leadership, we gain share because it is very TCO-based [Total Cost of Ownership]. In the client space, even when we have a better product, we may or may not gain share because there's a go-to-market side, and a developer side; that's the difference.

So, yeah, absolutely, we want to be the best [in the data center]. That’s why EPYC now has one-third of the world's market share.

On the PC side, we've had a better product than Intel for three generations but haven’t gained that much share. So, to me, that means that it's the developers, it's the go-to-market, and that's where I'm focusing now. I think building a great product in the client [consumer] market gets us to 20% market share by pure grinding, but to go to 40% is another gear, and that’s the machine I’m trying to build.

But don’t worry, I love gaming. When I present to the board, I say gaming is a strategic pillar in my strategy. I actually talk about a few things: commercial, PC, and gaming.

[..] Don’t worry. We will have a great strategy for the enthusiasts on the PC side, but we just haven’t disclosed it. We'll be using chiplets, which doesn't impact what I want to do on scale, but it still takes care of enthusiasts. [...] Don't worry, we won’t forget the Threadrippers and the Ryzen 9’s."

Now, this was immediately picked up on by various media outlets and through Tom's own analysis as being quite the move... Even, I, in my infinite wisdom, was found to be in awe of AMD's apparent turnaround from their prior, previously held conclusion that top-end products drive sales in the lower-end, too!

The question we're facing is whether this is actually true or not!

I would argue that, on a global scale, marketing works on this entire principle. You see a pro athelete or team wearing the products of a certain brand and you will be more likely enticed to buy something from that brand in order to "attune" or "align" with your favourite or aspirational vision of reality.

In clothing, fashion, and general societal cachet, these assumptions are met with a general agreement - whereby consumers will attach themselves to ancilliary products attached or aligned to their aspirational models.

In gaming? For console gamers this model works well; with influencers depitcting the "image" of "ideal" that the consumer would be able to aim for. 

For PC gamers, I would argue that this is still highly prevalent but to a lesser degree when it comes to hardware. My reasoning is that PC gamers, for the most part, have some technical knowledge, meaning that they can be more discerning of the merits of what they wish to buy to persue their hobby. 

Conversely, many sports fans do not know how to sew a shoe or shirt. This limits their participation to a level of "support" or "no support". Now, a sports fan's support may arrive or be applied on many different levels - from buying beer and watching events in a sponsored bar, or all the way up to commisioning personalised branded items from an accredited outlet.

Gamers don't have as many levels of "appreciation" available to them and often, "gamers" as a percentage mass of society, do not coalesce into such such a monolithic society of appreciation as sports fans do and are able to...

Gamers want performance and lack of technical issues. They have a more direct input into their own experiences and executive decisions than the typical sports fan and that needs to be captured in any analysis of the market other than a simple "winner affects all purchasing decisions rhetoric"!


A buyers market...


The Reality...


These comments may not actually reflect AMD's position on the client GPU market in the way we think they do. Mr. Hyunh interchanges CPU, GPU, client and business throughout his brief blurb. It's very difficult to pin down any particulars to the specific market segments other than the fact that "volume" and "marketshare" is the priority in the consumer (client) GPU space.

That doesn't preclude high-end parts. It doesn't preclude any specific products. What it does, or more specifically, what he says is that he is going to focus on developers to help move products.

That's the big takeaway.

To put it another way, my re-reading of this interview is: "Relationships with various business segments will drive user adoption."

In that context, the interview is a big nothing. Who (or what corporate entity) wouldn't state that they plan to grow the business by focussing on the business relationships which create value with which to grow the total addressable market for those same products?!

I think this interview and these quotes are taken completely out of understanding where the speaker is in their mindset.


The Hope...


There is one thing that I see in this interview that I am positive about and I've not seen it mentioned anywhere else: AMD may spend the money that they would have spent on a hypothetical high-end product on developer buy-in and adoption.

All these years, Nvidia have been spending away on "RTX" implementations to help developers use those technologies and, by and large, AMD has sat on the sidelines ignoring the issues developers face in the short-term, instead focussing on what developers face in the long-term. The problem is that consumers don't care about the difficulties of developers and will never worry about back-end problems. 

What they will care about is features added to new games that they can use to get more "performance" (read: usability) out of their hard-earned hardware. Nvidia offers that, AMD doesn't - or at least not to the extent of Nvidia.

While I disagree with the concept that AMD don't need to compete at the high-end of the performance spectrum to "enable" lower end buy-in from consumers. I do see the need for them to spend more of their resources on enabling developers to push their implementations and hardware to help their public perception and sell more hardware units. If this can't be achieved with their current budget considerations, then pulling away from one-upmanship at the premium tiers to focus on this for their architectures is a huge win, in my opinion...

This strategy doesn't preclude them from re-entering the market with earth-shattering performance at any future time and date. It just tells you that they want to become synonymous with features, like Nvidia are.

AMD are already comparable in terms of stability and performance. What they lack is that extra mindshare, both in the sellers and consumers.


Finale...


If, all things being equal, Nvidia was no more of a choice over AMD than a monitor with "Freesync with G-sync compatible" hardware then it is conceivable that the hardware share of the sales market* would increase for AMD... if they were cheaper.

Despite any business-orientated hurdles that the company may have to overcome, the final hurdle will be user-mindshare and that, unfortunately, is very much in the minority. To turn that perception around, AMD will need to compete in the higher-end space. It's simple psychology...
*Because that's what we're talking about here...

2 comments:

DavidB said...

To take market share from nvidia, they can't compete solely on price, to do so gains no mind share. AMD needs killer graphics features (and working with game devs to implement them) that don't exist on nvidia or intel.

Duoae said...

This is one of the points, yeah. They need to take that money that would have gone into developing the higher tier products and invest it elsewhere, like updating or adding software features...