1. If you have capital to invest, you could do worse than AI startups at the moment.
2. Nvidia's long-term threat is not just direct competitors (AMD, Intel), but the big cloud-players going to in-house chips. Supporting the next wave of your customers makes sense.
3. Using Nvidia is the path of least resistance right now. If you only invest in startups using your products (and you are an active investor), you give startups another reason to avoid taking a risk on the alternative.
I don't take it as this at all. I think NVidia is assuming (like the rest of us are) that AI is going to dominate the next decade or so of computing (at least?) and by investing in essentially _all_ of the independent players, they stand a good chance of backing a winning horse.
It almost turns them into VC firm more than market manipulators, imo
I think your Levi’s analogy is underselling it. I don’t know what % of gold income went to buying pants but I doubt it’s as high as an AI startups GPU costs. It’s really interesting that NVidia could be getting a huge chunk of this investment money back as revenue. This investment plan would probably be profitable just from driving up valuations and getting other investor’s money as sales revenue even if their equity positions go to zero.
Really though, they probably invest by giving GPUs or future-looking GPU commitments at-cost so they get equity at a discount equal to whatever margin they have in their chips.
> Ounce for ounce, a wafer of a top end NVIDIA chips is the most valuable substance on earth.
I think that would still be Californium. Even if NVIDIA could fit 300 H100 on a 300mm wafer at $40,000 a GPU a wafer would be worth $12,000,000 which would be in the same ballpark as a single gram of Californium.
Technically anti-matter would cost orders of magnitude more but there's no demand for it so it's not really valuable, whereas Californium is tightly rationed for scientific research. </silly-nitpick>
I don't agree with the GP, but I tthink the logic is that most of the money that AI startups pay is ultimately going to Nvidia. So, if most of the money is coming from Nvidia and going back to Nvidia, then Nvidia is self-dealings to some extent and artificially increasing their own valuation.
Of course, this doesn't make it a Ponzi scheme, it would at best be a different type of fraud. And it ignores the fact that the money isn't moving so directly - the AI companies are typically buying GPU compute power from a cloud provider, and it is the cloud provider paying Nvidia.
Still, if you're investing in Nvidia, you should keep an eye on this. It may mean that some of the increased demand for Nvidia GPUs is artificially sustained by Nvidia itself.
Presumably, this person thinks so because a lot of the cost of running an AI company is buying GPUs, which is mostly going to NVidia. The implication with this headline is that NVidia is propping up sales or driving demand by giving money to startups who will combine it with other investors money and eventually pay it back to NVidia.
Is it a Ponzi scheme? Not strictly, but is interesting to see a company investing in their customers. There’s nothing wrong with doing so, but it’s interesting to see it at a time when demand is so high it can’t be sated.
People think it is a problem because with AI startups most of the money in these investments comes right back to nvidia as revenue.
So nvidia gets to do this:
1) start with $100 M
2) invest in startup
3) get investment worth at least $120 M (investment at discount)
4) reflect 20% gain on books
5) receive $80 M in guaranteed revenue (H100s)
So before the inevitable long-term failure of 90% of these startups, nvidia gets to reflect a nearly immediate 100% gain on any money they run through this.
Then they use the gain to invest in more startups where the only thing guaranteed is the purchase of nvidia chips.
When the startups go to zero and the market is flooded with used H100s they probably believe they will release a new arch and repeat.
It's an investment/gamble that is realized as revenue now, but is cashflow negative. This may be competitive with throwing that money at additional sales/marketing activities, but long term creates the risk that they have to write off a lot of equity and the whole program will have been a loss.
Cashflow shouldn't be affected - might even be positive(?). If they invest $X in a startup, it reduces the cash asset but they'll have an offsetting "other investments" based on the valuation of the startup - which should be at least 100% of the total investments the startup has received. I'm guessing here, but, there's a good chance that it'll be cash flow positive as is.
Not to mention the fact that most of the money from this investment flows back to Nvidia, adding to the cash flow as revenue.
1. If you have capital to invest, you could do worse than AI startups at the moment.
2. Nvidia's long-term threat is not just direct competitors (AMD, Intel), but the big cloud-players going to in-house chips. Supporting the next wave of your customers makes sense.
3. Using Nvidia is the path of least resistance right now. If you only invest in startups using your products (and you are an active investor), you give startups another reason to avoid taking a risk on the alternative.
edit: typo