News AMD 1Q2024 Earnings

[maddie]

Fabricated Knowledge just dropped an article on AMD’s earnings. And it’s a big OOF. The MI300 family only has 2-3 quarters to shine till Blackwell’s B100/B200/GB200 puts it into the ground and expectations on the buy side we’re that they’d get 5-6 billion in 2024. So they really aren’t getting a lot of the profits out of the AI boom as Nvidia and Broadcom who does the chips for Googles TPU’s, Amazon’s Trainium and Inferentia processors and possibly Microsoft’s AI chip. They’re in a distant third place behind Nvidia and Broadcom but thankfully are ahead of Intel’s Gaudi 2/3 sales. Also if they don’t manage to get the MI400 out ahead of the R100 that’s going to get announced next year it’s going to be a dud in sales. https://www.fabricatedknowledge.com...lers?r=reiiy&utm_campaign=post&utm_medium=web

Well,,,,,,,,,,, it's fabricated knowledge after all, so be cautious.

Could. Not. Resist.

 

[Abwx]

Well,,,,,,,,,,, it's fabricated knowledge after all, so be cautious.

Could. Not. Resist.

That s actually fabricated since he states that AMD has 2-3 quarters left before the competition is RAMPING its next product in Q4, as if ramping was releasing something, let alone delivering.

 

[Mahboi]

AMD stock taking it on the chin today, down nearly 10%. Seems AI hype deflation causing some investors to seek other options.

I wouldn't put much credence on those.
I'm convinced that the vast majority of those shifts are all pre-programmed bots.
"Wait for date X hour X, if stock hasn't gone up by x amount, sell".

Pre-buy the stock, wait for the big announcement that'll make the prices go up, if they don't, sell at roughly the same price, if they do, hold and wait for the human owner to decide to sell or not.

 

[Mahboi]

For client in general: it doesn't matter if it is much faster Intel will spam cheapo RPL laptops at slightly lower prices and it will sell better still.

How much lower do you think Intel prices can get? Have you seen their earnings? It's not possible.

For DIY having the fastest gaming CPU means certain people will buy it no matter what. But it isn't a huge market.

ST is still King, despite what anyone says.
I wonder how the server market will take it. I remember this statement from Amazon's VP that "Compute wasn't the problem" for them a year ago.

 

[Hitman928]

Fabricated Knowledge just dropped an article on AMD’s earnings. And it’s a big OOF. The MI300 family only has 2-3 quarters to shine till Blackwell’s B100/B200/GB200 puts it into the ground and expectations on the buy side we’re that they’d get 5-6 billion in 2024. So they really aren’t getting a lot of the profits out of the AI boom as Nvidia and Broadcom who does the chips for Googles TPU’s, Amazon’s Trainium and Inferentia processors and possibly Microsoft’s AI chip. They’re in a distant third place behind Nvidia and Broadcom but thankfully are ahead of Intel’s Gaudi 2/3 sales. Also if they don’t manage to get the MI400 out ahead of the R100 that’s going to get announced next year it’s going to be a dud in sales. https://www.fabricatedknowledge.com...lers?r=reiiy&utm_campaign=post&utm_medium=web

He makes some good points, most of which have already been discussed in this thread, but his negative outlook for AMD seems to hinge mostly on 2 things. First, that MI400 will be a dud, and second, that the x86 market is dying and that Intel will give them increased competition soon in said dying market. Can't say that I agree with either of those takes. Maybe he ends up being right, but he doesn't really provide evidence to support them so it seems to just be him speculating that because AMD isn't completely supply constrained this year, that no one will want their next gen stuff (completely ignoring that AMD MI sales went from $0 to $4B in sales in 12 months).

 

[Hitman928]

I wouldn't put much credence on those.
I'm convinced that the vast majority of those shifts are all pre-programmed bots.
"Wait for date X hour X, if stock hasn't gone up by x amount, sell".

Pre-buy the stock, wait for the big announcement that'll make the prices go up, if they don't, sell at roughly the same price, if they do, hold and wait for the human owner to decide to sell or not.

Post earnings stock movement tends to be very volatile around growth stock companies. The stock has already regained around 75% of what it lost right after earnings so it seems the more long term investors are sticking around.

 

[Mahboi]

Fabricated Knowledge just dropped an article on AMD’s earnings. And it’s a big OOF. The MI300 family only has 2-3 quarters to shine till Blackwell’s B100/B200/GB200 puts it into the ground and expectations on the buy side we’re that they’d get 5-6 billion in 2024.

2-3 quarters?
So you mean before MI350X or MI400 comes out?

So they really aren’t getting a lot of the profits out of the AI boom as Nvidia and Broadcom who does the chips for Googles TPU’s, Amazon’s Trainium and Inferentia processors and possibly Microsoft’s AI chip. They’re in a distant third place behind Nvidia and Broadcom but thankfully are ahead of Intel’s Gaudi 2/3 sales.

Sounds like raw copium to me.
Especially comparing it to Gaudi that it would "thankfully" beat. It trounces it.
AMD wasn't going to displace Nvidia in one go, especially before ramping up production. If they get even 15% of the AI market on their first big product, it's already a solid start.

Also if they don’t manage to get the MI400 out ahead of the R100 that’s going to get announced next year it’s going to be a dud in sales. https://www.fabricatedknowledge.com...lers?r=reiiy&utm_campaign=post&utm_medium=web

More raw copium. That guy has no idea what MI400 has and is assuming that R100 is going to be the best thing ever.
It sounds like he's following the Nvidia marketer handbook to the letter frankly. Not a single reason to buy his drivel.

 

[Mahboi]

Post earnings stock movement tends to be very volatile around growth stock companies. The stock has already regained around 75% of what it lost right after earnings so it seems the more long term investors are sticking around.

Yes you don't give the entire bank to the bots, just the money you're willing to only mildly control.

 

[Ghostsonplanets]

Fabricated was hyping up and gloating over B200/B100 on Twitter after the announcement. Even if he raise some salient points, I can't shake but think his views are tainted by his Nvidia leanings.

 

[Mahboi]

I read that as "Nvidia learnings" as in he did his classes on tech at Nvidia, marketing section.

You'll always find sycophants claiming that AMD is "disappointing" even when they launch a juggernaut obliterating anything Nvidia has.
It's always the same pattern too:
- AMD can't do what Nvidia can
- AMD did better than Nvidia, but it won't work
- It works and sells but it won't sell for long
- It won't get replaced by something better, Nvidia best
- It may have gotten replaced that time, but nothing guarantees that AMD can do it again

And so on. NV GOD or denial.

 

[Abwx]

Fabricated was hyping up and gloating over B200/B100 on Twitter after the announcement. Even if he raise some salient points, I can't shake but think his views are tainted by his Nvidia leanings.

That sound like the typical guy who was defrauded of his money by buying both Nvidia and Intel s stocks on the highs.

Since i know no one who would be ready to waste his time enriching the others it look like he s just trying to lure the unsuspicious buying those firms stocks as a mean to get back in the money.

 

[adroc_thurston]

Fabricated Knowledge just dropped an article on AMD’s earnings. And it’s a big OOF. The MI300 family only has 2-3 quarters to shine till Blackwell’s B100/B200/GB200 puts it into the ground and expectations on the buy side we’re that they’d get 5-6 billion in 2024.

He has no clue.
Expected, he's dolan-adjacent.

 

[Mahboi]

https://www.reddit.com/r/AMD_Stock/comments/1cl4w66/amds_mi300_disappointment_hyperscalers_capex_and

Even Plebbit is calling his statements out.

 

[DAPUNISHER]

The Mercury Research report for AMD CPU sales is good -


https://www.tomshardware.com/pc-components/cpus/amd-takes-cpu-market-share-from-intel-in-desktops-and-servers-but-intel-fights-back-in-laptops

 

[gdansk]

2024Q1 Server: 23.6% unit share but 33% revenue share.

That's the best segment for them.

 

[adroc_thurston]

That's the best segment for them.

about to get a whole lot better

 

[adamge]

about to get a whole lot better

We've been saying that for 5 years and they've only made modest progress.

 

[adroc_thurston]

We've been saying that for 5 years and they've only made modest progress.

Over 1/3rd revenue share is a lot in slow markets.
Why the cope?

 

[adamge]

Over 1/3rd revenue share is a lot in slow markets.
Why the cope?

I just don't see a tipping point happening in the near future. The performance will be better, but it's always been better. The efficiency will be yards better, but it's always been yards better. So where does the jump in market adoption come from all of a sudden?

 

[adroc_thurston]

So where does the jump in market adoption come from all of a sudden?

Specific workload performance for enterprise is the key Turin selling point.

 

[KompuKare]

2024Q1 Server: 23.6% unit share but 33% revenue share.

That's the best segment for them.

Intel are not quite wrapping Xeons with dollar bills like the did with dumping (er, "contra revenue" in Intel-speak) when they tried to get into the tablet market with Atom (far too late and in a great panic), but while:
~ 75%+ for 66% of the revenue
doesn't sound that bad for Intel, I suspect that in certain market segments they must be giving Xeons away, or at least once you take into account their fixed costs.

 

[Ajay]

Specific workload performance for enterprise is the key Turin selling point.

Enterprise sales are picking up? That's interesting.

 

[adroc_thurston]

Enterprise sales are picking up?

yea Lisa said as much on recent call.

 

[Tuna-Fish]

Intel are not quite wrapping Xeons with dollar bills like the did with dumping (er, "contra revenue" in Intel-speak) when they tried to get into the tablet market with Atom (far too late and in a great panic), but while:
~ 75%+ for 66% of the revenue
doesn't sound that bad for Intel, I suspect that in certain market segments they must be giving Xeons away, or at least once you take into account their fixed costs.

I don't actually think that's true, it's just that the segments of the server market they are still holding are the ones with the lowest ASP.

At this point, AMD is basically dominating hyperscaler sales. If you are purchasing mainly the highest-end chips, and are doing so by the truckload, you are probably already getting more than half your x86 product from AMD. In contrast, Intel is holding firm in traditional enterprise sales, selling really well to companies whose primary business is probably not computing, but that need to purchase servers to support whatever it is they do. Intel's strength here is existing sales channels and relationships, and the fact that in this market no-one is really doing performance comparisons between individual chips, or even really know what those are, they are just going to middlemen with problems and getting them solved.

yea Lisa said as much on recent call.

It's improving, with high % growth but from a low base. Gonna take a while yet.

Specific workload performance for enterprise is the key Turin selling point.

I honestly doubt performance is going to move the needle that much, they already have that. The big difference is more about building relationships with the vendors that actually sell to these customers, and it's ongoing but inherently slow.

 

[adroc_thurston]

I honestly doubt performance is going to move the needle that much, they already have that.

No-no, not quite, they don't.