Discussion AMD Earnings Q3 2025

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
J

Joe NYC

Diamond Member
Jun 26, 2021
3,854
5,386
136
adroc_thurston said:
Not much, even this Q EPYC grew more than Instinct.
Click to expand...

Some comments from the earnings call re: EPYC were interesting:
- Lisa said that the shipments have transitioned to 1/2 Turin
- AMD shipping early Venice silicon already, and she said in Venice generation, ramp will start quite rapidly right from the announcement day.

In the meantime, Intel said they are having trouble shipping enough 10nm server CPUs (SPR, EMR).

This indicates that AMD is continuing to take the high end and intel is still stuck on low end.

Also, if Venice is getting such a positive reception, it could be on its own merits, but it could also be reflecting expectations for Diamond Rapids.
 
  • Like
Reactions: Tlh97 and lightmanek
adroc_thurston

adroc_thurston

Diamond Member
Jul 2, 2023
7,580
10,345
106
jpiniero said:
Wall Street doesn't care about Epyc. Only AI.
Click to expand...
Yeah they do, everyone noticed that CPU capex is alive and well.
Joe NYC said:
This indicates that AMD is continuing to take the high end and intel is still stuck on low end.

Also, if Venice is getting such a positive reception, it could be on its own merits, but it could also be reflecting expectations for Diamond Rapids.
Click to expand...
DMR is a non-factor until late 2027.
Joe NYC said:
That is true. It seems WS is holding it against AMD that all their other divisions are performing so well...
Click to expand...
It's all fine.
 
G

gdansk

Diamond Member
Feb 8, 2011
4,679
7,896
136
Thunder 57 said:
And their market cap is about 1/6 of the US GDP last I checked. A lot of people would lose a lot of money if/when the AI bubble pops.
Click to expand...
Market cap and GDP is a somewhat silly comparison even if they both have dollar figures.
What portion of the S&P 500 are heavily AI exposed might be a better metric for AI bubble valuation.
 
  • Like
Reactions: Thunder 57
Thunder 57

Thunder 57

Diamond Member
Aug 19, 2007
4,130
6,881
136
gdansk said:
Market cap and GDP is a somewhat silly comparison even if they both have dollar figures.
What portion of the S&P 500 are heavily AI exposed might be a better metric for AI bubble valuation.
Click to expand...

Yea but that's more work, unless, I ask AI to do it! And of course it wouldn't be just Nvidia that would lose out, a good bit of other big players would as well. NVDA is just easy to pick on because it's the largest and if that 90%+ revenue from AI is true that is hardly diverisfied.
 
J

johnsonwax

Senior member
Jun 27, 2024
438
651
96
Panino Manino said:
How big will be the blow on AMD?
Will AMD lose most of the value it got in recent years?
Click to expand...
So for folks wondering about this stuff, some lessons from an investors viewpoint from 2008.

If the AI bubble pops, everyone will get shivved. During big calamitous events, the 'weighing machine' is too slow to do analysis and everyone pulls their money. Sometimes there's a solid alternative for capital - bonds, commodities, real estate etc. but right now everything looks like dogsh!t so my guess is it'll pull to cash because the existence of that cash will be in question. Additionally the scale of this bubble is much larger than 2008. In 2008 there were questions about whether people would have places to live, jobs, and so on. Fundamental questions like 'who will have the money to buy a car' were being asked. That will happen again because nobody is really clear on what the first order effects will be - which financial institutions are over leveraged into AI, let alone the 2nd and 3rd order effects. In 2008 you had money market firms failing because they could convert equities into cash fast enough to pay depositors asking for their money - firms with zero exposure to mortgages. Mainly that was a lack of liquidity. Everyone had debt against assets and nobody had cash.

In the days after as the scale of the calamity becomes apparent the weighing machine starts to work again and money starts to flow back in. Everything tech will look like crap, but you know, people will still have squeaky doors and WD40 has a nice dividend so let's put some money there, and the fundamental instruments of the economy will recover - the necessities - food, utilities, critical services like healthcare, etc. Then the more detailed questions get asked - what was AMDs revenue/profit exposure to AI? Do their customers have any money? Did any of their suppliers go under? AMDs business model gets completely re-evaluated and if they look isolated, they'll recover. My guess though is that being strongly dependent on server revenue, those are the folks that won't have any money and they'll take a longer term hit until demand returns. The real hit is going to be against the Mag 7, with Apple being the fastest to recover because they generate literally no revenue off of AI - for them AI is a cost center, not a revenue/profit one, and they are only exposed to consumer spending. For everyone else, they are wildly exposed to AI as a revenue center. We may find Amazon, Google, Microsoft all getting stripped back to the studs in a collapse and their ability to buy AMD silicon nonexistent, at least for a while. It's hard to tell where the fallout in datacenter will stop. Clearly it would nuke all AI datacenter, but it's likely to not stop there and take out a lot of conventional datacenter as well.

Also be aware that banks, housing and crypto are all likely to fail to some degree or another as well. These are all asset stores with very little available cash and everyone is going to be chasing the same relatively small pile of earth dollars. As such, there is unlikely to be sufficient mechanisms to keep them all from failing. In the GFC, you can't fully divorce the auto bailouts of 2007 from the housing crash from the bankruptcy of Bear Stearns. They are all connected, as was the millions of jobs subsequently lost, homelessness, the cascade of business failures that resulted from that, and so on. And that was at a time when the people running the US government actually mostly knew what they were doing. I have no idea what this crew who seem to lack any theory of the economy whatsoever might do.
 
  • Like
Reactions: adamge and lightmanek
DrMrLordX

DrMrLordX

Lifer
Apr 27, 2000
23,019
13,118
136
@johnsonwax

Seems like an over-dramatic take. Businesses that don't crash and burn immediately from a "pop in the AI bubble" will do fairly well, especially 2-3 years out. In fact people who had broad market exposure to the stock market in 2007-2008 had all recovered by 2010. The only ones that had suffered longterm losses were those directly involved in mortgage-backed securities or those who had held questionable mortgages prior to the crisis. People can act scared and attempt to grab cash, but the only ones who will win out with that strategy are those who manage to liquidate before markets crash and then buy back in near market lows.

AMD has MI450 coming up, which may be affected by a market downturn, but they have successful products elsewhere in the datacentre market that will still be in demand. They should be fine.
 
J

johnsonwax

Senior member
Jun 27, 2024
438
651
96
DrMrLordX said:
@johnsonwax

Seems like an over-dramatic take. Businesses that don't crash and burn immediately from a "pop in the AI bubble" will do fairly well, especially 2-3 years out.
Click to expand...
Yeah, but I'm going to note the context of 'recent years'. The recovery of many of these stocks will take equally as long, so depending on what your plans are for that investment - sell it to buy a house, send a kid through college, whatever - you might be on the sidelines for what you consider to be a long time.
DrMrLordX said:
In fact people who had broad market exposure to the stock market in 2007-2008 had all recovered by 2010.
Click to expand...
I think that's overstating things as it was highly dependent on where you were. If you were using that for retirement income and expecting to draw it down at 5% per year, you were now drawing it down at a much faster rate, meaning you had a lot less in equity during the run back up. A lot of retirees really got hosed pretty badly there, but if you were 5 years out from retirement, not so much. And you have to also account for the various secondary effects from that, yeah your portfolio might have recovered but we lost nearly a million jobs per month for half a year - and if you were caught up in that, your portfolio recovery might have been small comfort.
DrMrLordX said:
The only ones that had suffered longterm losses were those directly involved in mortgage-backed securities or those who had held questionable mortgages prior to the crisis. People can act scared and attempt to grab cash, but the only ones who will win out with that strategy are those who manage to liquidate before markets crash and then buy back in near market lows.
Click to expand...
Or were Bear Stearns investors or Merrill or Lehman or WaMu or Wachovia or UBS or Chrysler or Charter or Tribune or countless smaller firms that got acquired out of chapter 11, or are just too small for you to remember.

The problem with your take is that you are doing it with the benefit of hindsight. In the middle of it, nobody knew where the bottom was, how long it was going to last, whether you could hold long enough to recover, and so on. The fundamentals of the market changed significantly which put retail investors at a large disadvantage because the pros and insiders could see the landscape faster and better than the rest of us. How much of your kids college fund were you willing to risk in the event the government efforts didn't work? How much could you afford to hold onto after you lost your job and had no income for a year - or were you forced to sell after losing half its value? That's the hard part. What caused the market to fall? It wasn't a drop in revenue or profits, it was people selling their stock, even if it meant taking a loss. People were selling, and not all of them were in the market to ride it back up. And a lot of retail investors didn't have the excess capital to take the risk of HODLing - selling at 50% down meant they got to keep the other half, and for all anyone knew, they were at risk of losing that as well. Like I said, the fundamentals of the market shifted radically and in ways that normies couldn't see.

I was options trading during this period which meant that basically I never slept well and one day I'm sitting on a smallish long out of the money call option - maybe $4K or something. Not really a lot of money, but I was down about 50% overall and that $4K felt a lot bigger then. This was pretty close to the bottom, but I had no way of knowing that. I had my finger over the execute button and I chickened out. That would have been a >$1M payday had I done it. I have the hindsight now, but didn't then. The person on the other side of that trade would have handed me >$1M in stock in exchange for $4K something like 6 months earlier. A lot of money changed hands despite the fact that the markets recovered and most of that money didn't flow to retail investors. Most of it flowed to pros from the retail investors. In aggregate the 'market' recovered, but that doesn't mean everyone recovered or even every category of person recovered. You know who the fastest growing population of homeless are? Boomers, a lot of whom lost their homes and most of their portfolios in 2008 and lost what was left during Covid. I know a lot of people that were building nest eggs into that period and got wiped out. People like me, who could hold because I had a fairly safe public sector job got rich at their expense.

DrMrLordX said:
AMD has MI450 coming up, which may be affected by a market downturn, but they have successful products elsewhere in the datacentre market that will still be in demand. They should be fine.
Click to expand...
Apple had just launched the iPhone - the single most valuable product ever created - and lost 50% of their value. They recovered obviously, but the fundamentals of a company mean jack sh!t during a market panic and economic crisis. That returns, but it can take a while. You say the datacenter market will still generate demand. What makes you think they'll have any money to buy them when they are not divorced from the AI datacenter market that will almost certainly get annihilated? The housing crisis wasn't contained to the subprime lenders. It took out S&Ls, builders, retail property investments when communities of homes emptied out and there was nobody to go to the restaurants. The construction industry basically stopped for a few years. I know construction project managers that had no work for years. There's always a ton of secondary and tertiary impacts from these things and quite often they're the demand you're relying on now and they may not exist in the immediate aftermath. And I'll note the amount of misallocated capital right now is something like 5x larger than it was in 2008. Overall the economic conditions going into 2007 were pretty good. They aren't now, they're actually pretty terrible right now. The scale of this crash is potentially going to be much larger.

My bigger point is that if you are thinking 'are AMD investors exposed to an AI crash'. F'ing Burger King is exposed to an AI crash. The entire economy is exposed - at least for a year or possibly several depending on how competent the government intervention in and whether the government can even intervene on this scale. When the feds agreed to throw less than a trillion dollars into the entire financial sector to provide liquidity, that was considered an impossibly large number at that time. OpenAI - just OpenAI - has asked for a $1.4T backstop. Now add in Oracle, Nvidia, Google, Microsoft, Amazon, and so on. The US may not be capable of shoving a floor under such a crisis.
 
  • Like
Reactions: Kryohi
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom