oh ok then I'll just give up on hoping things change, thanks for saving us all the time

It's clear that American capitalism and AI are now fully intertwined. Either they both collapse or they both succeed. This bubble is actually too big for the government to bail out and if it tries it'll probably just bring the American dollar down with it, like tying helium balloons to the Titanic.

Be careful, rsoto2, it sounds like you're at risk of anthropomorphizing the lawnmower. [1]

[1] https://www.youtube.com/watch?v=-zRN7XLCRhc&t=2308s

All the people who ever worked on Oracle iProcurement fired into a volcano?

Natives collectively possessed all of Hawaii long before the concept of private ownership arrived.

Well over? It was taken by force about 130 years ago

A whole century? The word "never" is doing a lot of work here.

Oh they certainly could default. They make $70B a year in revenue and $17B in profit. Their current interest payments are $5.7B a year!! Their debt is only 1 grade above junk.

But the truest measure of their potential to default, their credit default swaps, are pricing in a 20-22% chance to default before 2031.

What is their core business today? I remember when it was the enterprise database server product, but I can't imagine that is true anymore.

Claude, build me a copy of Oracle RDBMS. Make no mistakes

I would be very very curious in some source of these Nvidia buybacks

> you only need to replace the compute board

"Only" is doing a lot of heavy lifting here, considering that this compute board makes up the majority of the cost. And if they start going with bigger designs, like they’re doing now (compute in memory, custom TPUs etc), then even this won't hold.

> If you want to do an upgrade that increases power use, you may have to get a new rack.

Blackwell to Rubin upgrade is exactly that case. 3x3U 110 kW units vs the 8×1U 33 kW shelves typical in GB300 NVL72, power whip rating rises to 100A from 60A, rack coolant flow doubles versus GB300, and both compute and NVSwitch trays are fanless, 190-230 kW per rack depending on power mode vs about 140 kW for GB300 NVL72 and NVLink switching is new too etc

And where is your source about buyback program? I found no Nvidia trade in or buyback program for data center hardware. Used gear is resold through third parties, ITAD vendors like Procurri and Bitpro buy and resell used GPUs from hyperscalers, and resellers like Alta Technologies sell refurbished units with warranties.

The closest real thing is a financing arrangement announced in August where Nvidia partnered with Apollo, BlackRock, Blackstone, and others and in SOME CASES may provide residual value support for up to 25% of an opportunity, assessed project by project. So in practice, if a borrower defaults and the lender has to liquidate GPUs that can't fetch their booked value, Nvidia covers part of the shortfall. That backstops lenders. It is not a buyback for customers who upgrade.

What does Nvidia do with the parts they buy back? Resell to others or grind to dust?

I'm not really sure they have thought this through. Or at least their solution has been to essentially insulate the company and even the physical building from the debt used to buy the GPUs. These SPVs typically lease the building and secure the debt to the GPUs themselves so if it all goes pear-shaped, it'll be the debtholders who are left holding the bag, priamrily.

But I have questions:

1. How often do GPUs fail? I've seen figures that this is as high as 9% annual failure rate for DC GPUs;

2. How easy is it to get replacement stock? At what cost?

3. If you trade in, what value do you get for your GPUs?

4. What is the availability like? How does NVidia decide who gets the next wave of GPUs? That's NVidia's big power here. It's a bit like how Boeing and Airbus decide on customer order and if you upset them, you may lose other orders. This happened to Qatar Airways over the A380 [1].

5. How long are the current form factors and interfaces going to last? How long before the power requirements get larger than what DCs are currently built for? Blackwell GPUs use a lot more power than, say, 2 or even 1 generation ago did.

GPUs are a huge cost factor, allegedly more than 50% of the cost of a data center [2]. If you take a bath on the trade in and are forced to do it because you can't compete with the current generation's performance-per-watt, that's not a good situation to be in.

[1]: https://www.arabnews.com/business/qatar-airways-in-mystery-a...

[2]: https://orennia.com/insights/what-it-actually-costs-to-build...

The real shame with all of this is that when (not if) the AI bubble pops, it's going to be like the 1990's, when there was a huge supply of unix workstations, etc.

The difference is that none of this Nvidia gear can run on normal 120V (or even 240V) circuits, so it'll end up going to the landfill or just being powered down in place.

There is no plug-and-play upgrade SKU that allows you to drop Rubin compute boards seamlessly into a Blackwell rack. A Blackwell NVL72 rack pulls around 100 kW to 140 kW. A Vera Rubin NVL72 rack is estimated to draw up to 190 kW to 230 kW per single rack. Rubin is also liquid-only cooling. Further it uses NVLink 6 which requires a different backplane.

Where is there evidence of buyback program for Blackwell boards?

Commodity products can't demand high profit margins. Supermarkets typically have 1% to 3% net profit margin. SaaS often has 70% to 85% profit margins. Data centers hover around 15%. You don't want to sacrifice a high profit margin business in order to gain share in a low profit margin one.

It's really telling how they just assumed all the permits would be approved, putting in no effort to ensure that would actually be the case. Because until now they've always been rubber-stamped on anything they do.

if living good was the goal post for Larry he'd have quit decades ago.

Curious to know why you think ERPs are done. I actually do somewhat agree with you in that their best days are behind them.

It is going to be incredibly difficult to bill outrageous amounts for armies of consultant to implement or customise the product when agents can do the same for much less, or at-least a 5-10x reduction in headcount for similar complexity of changeset.

If anyone is in the industry, please feel free to chime in as to what you are seeing on the ground.

Yet, entangle they did: https://openai.com/index/five-new-stargate-sites

> In July, OpenAI and Oracle entered an agreement to develop up to 4.5 gigawatts of additional Stargate capacity. This represents a partnership that exceeds $300 billion between the two companies over the next five years.

> their own AI services

Oracle doesn't have any. And that's wise. Oracle makes good RDBMS servers. They suck at everything else.

All AI players are in bed with each other, they have commitments and obligations to each other. One of them failing (namely Oracle) could bring them all down domino-style. And us with them, since the market is ride or die with these companies.

Oracle not able to provide its usual "services" would disrupt many institutions which unfortunately rely on their technology and assistance operating the often expensive and complex Oracle integrations they've set up, for better or worse. I really despise Oracle and the Ellisons and they don't need a bailout but I'd hate to see what could come from many significant collapses of technology for social institutions.

It’s literally dilutive. Current shareholders now only own 97.5% of the equity instead of the 100% they owned previously.

https://en.wikipedia.org/wiki/Stock_dilution

I'd like to note that "the sector which employs you" should be considered when talking about diversifying. (In general, not only when worried about the Robot Takeover.)

A long stretch of unemployment can be weathered by selling a bit from strong investments, and a crashing stock-market can be waited-out by staying firmly employed, but both simultaneously is what can force people to "sell low" just to pay regular expenses.

The dotcom crash has nothing to do with the AI build out

I think they are actually underpriced relative to "taking away" jobs. I personally believe ant at 1-2T is pretty reasonable if you consider claude to effectively be the "best devtool of all time," aka it literally writes all software for you.

I think (somewhat arbitrarily), to get to higher valuations, you are going to have to disrupt more than just the software business. Which is what most labs are trying with finance, biology, chemistry, law, etc. That might be enough for a few hundred billion in TAM more.

I think to get to a ~5T+ valuation, then we have to start thinking about AI meaningfully displaces classes of jobs, maybe in the single digit %'s. As in, _global_ SWE headcount starts going down, but revenues for software companies as a whole go up.

I think for ~10T+, you definitely need meaningful job displacement across most white collar sectors.

What a punishment, free up your days to spend your billions...

As an aside: that's just the problem: we don't anthropomorphize Larry Ellison, or his ilk.

If we did anthropomorphize people like him, he'd be in a lot more trouble than he is. Instead we legally and socially treat him like a lawnmower.

LOL..