AMD paid off $750M in debt this quarter and had +$439M in free cash flow.
Last edited:
Commercial volume.What exactly would the bull case be though
Newsflash: AMD hasn't won much commercial slots. Just yet.Corpos don't seem to care
No, it's because they run a fairly expensive roadmap.in fact the bad margins are likely because they simply aren't getting high enough prices.
Commercial volume.
It's beat and raise, what's not to like?
See, the $AMD story is simple.
If ML capex craters, the dollars would go back to server CPU spend and guess what AMD sells also.
Because they run a Big and Expensive roadmap that hedges on AMD winning commercial volume in the coming years (and they will, afaik).
I wonder how much money they are accounting for the increased RMAs because of the recent issues.
This is the n-th round of layoffs under Pat and it's definitely driven by their huge capex and low FCF.
The most affected upper SKUs aren't particularly high volume.
Volume drivers like RPL-U 282 are safe.
Gaming should also improve if PS5 Pro is releasing later in the year.
Who would have thought ~3-5% of market share would be so profitable. That AI bubble sure is big.
What I mean is, what exactly is going to change that will compel Corpos to switch?
That's what we're about to find out!What I mean is, what exactly is going to change that will compel Corpos to switch?
They're doing it iteratively, Intel issue is profits drying up due to DCAI meeting an untimely end.Almost hard to imagine the scale of CapEx needed to catch up to be competing with TSMC
Yeah but they're still gonna be solid enough.Margins on those are likely going to decrease, considering how far behind the leading edge these SKUs are going to be finding themselves behind.
Almost hard to imagine the scale of CapEx needed to catch up to be competing with TSMC. TSMC has been making these investments for over decades, and Intel planes to catch up in span of a couple of years.
Pretty sure MS isn't selling their Azure AI options at a loss so they are making money on it too.Not bad. They might be the only ones actually making money on AI besides nVidia.
TSMC was able to make those investments because of revenue from customers. Intel will get the commitments from customers if they can deliver on the roadmap. Even merely catching up to TSMC would be good enough to land a lot of business from customers who don't want the risk of having all their eggs in the Taiwan basket.
You certainly don't need "decades" of investments - it takes about 3-4 years to go from empty field to an operational leading edge fab if you already have the process roadmap which they do.
Unfortunately for them not everyone is like Apple with a giant checkbook to prepay for chips to fund future capital investment like they have with TSMC, but there's a ton of private equity money sloshing around right now since they're gunshy about real estate at the moment.
If I was Intel and had some big customers potentially interested but stymied by the "chicken and egg" situation (i.e. they don't want to commit to buying a lot of wafers from Intel without knowing for sure that the capacity will be online, Intel not wanting to build the capacity without knowing they have a long term commitment from a customer) I'd try to work some sort of three way deal with private equity. They fund the fab construction in exchange for say 8% interest, and Intel pays them back with the revenue - and the customer offers convertible warrants to the PE fund exercisable in the event they don't fulfill their contract to buy the wafers.
Intel just needs to get the ball rolling, once they get big enough they can continue the buildout using their own cash flow, or they can land Apple in the early 2030s and get prepayments and suddenly they're ahead of TSMC (if they lost Apple they'd lose 1/4 of their revenue, and Intel would gain it, so that covers more than half the distance between the two in one fell swoop)
Yeah. The whole "a fab is a fab is a fab" argument is way too simplistic.It's not as simple as having one production line or having one fab and you are there, you tied TSMC.
The talk the last few months has been that the silicon is already being stockpiled but it's not expected to sell in normal console numbers, the biggest dip is Xbox quarter by quarter collapse.But if AMD forecast of double digit decline in gaming is correct, there will be no PS5 Pro.
Another possibility is that Sony has not given AMD the final go ahead to start manufacturing the silicon. Silicon design has apparently been finalized a while ago.
The talk the last few months has been that the silicon is already being stockpiled but it's not expected to sell in normal console numbers, the biggest dip is Xbox quarter by quarter collapse.
Anyone interested on reading Ian Cutress substack about this?
It's here: https://morethanmoore.substack.com/p/amd-2024-q2-financials-its-all-ai
There's a transcript of the Financial Analyst Q&A.
I'm not following. Sales of consoles do decline as they age (unless some killer games come along).If 5 years into the console cycle your gaming segment is still getting obliterated by "lower PS5 sales", maybe it is time to reconsider the long term prospects in that segment.
It's the noegaems year, console decline is expected.maybe it is time to reconsider the long term prospects in that segment.
If 5 years into the console cycle your gaming segment is still getting obliterated by "lower PS5 sales", maybe it is time to reconsider the long term prospects in that segment.
That was in the last gen, not this gen.XBox and PS saved AMD from bankruptcy. Never bite the hand that fed you
If you lose 59% revenue primarily due to a product which already has a low base in terms of contribution, then that segment needs an overhaul. And FYI Radeon sales were up sequentially as per the transcript, not YoY.I'm not following. Sales of consoles do decline as they age (unless some killer games come along).
But AMD made many billions in revenue from IP they were developing anyway. They shouldn't pursue that market because its sales go down until the next generation? The gaming group still has a higher operating margin than the client group despite that. The PS5/XBX has been some of the easiest money AMD has come by.
Perhaps your quotes were to render the claim suspect but AMD said RDNA sales are actually up YoY so what else could it be?
What are your suggestions? I don't get it, really. They ought not sell IP to Sony and Microsoft in the future because revenue went down in the 4th year? I really don't see any other solution that makes Radeon group more money for the same R&D expenditure.If you lose 59% revenue primarily due to a product which already has a low base in terms of contribution, then that segment needs an overhaul.