I see that hot on the heels of the OpenAI/Oracle/Nvidia deal that I wrote about yesterday, Nvidia has demonstrated that it already knows that there is no money to pay Oracle to build data centres with its GPUs, so it taking a stake in OpenAI, reported at ‘up to’ $100B.
An approach along the lines of – ‘you’re ordering our stuff through a third party. but we know you won’t be able to pay for it, so here’s enough dosh to get the process rolling. Some of your fancifully valued stock in return would be nice!’
The problem with that is Nvidia will presumably build equipment worth the money, but the shares it is getting in return (‘a major stake’) are based on a valuation of OpenAI at this rather fairy tale moment in the market concept of accurate pricing. This deal is probably made of necessity (if the whole mega project is to move forward) but looks like a potential disaster for Nvidia to me.
An aside and something not explained in the press release is that OpenAI is not currently a public company as I understand it, so I am not sure of the corporate governance implications.
Nvidia’s much smaller but likely more accurately valued investment in Intel looks a lot more prudent. At least they are diversifying while splashing the cash.
However, I will be nervous about funds that have Nvidia well up in their top 10 holdings from this point onwards. The puff of fresh air into the bubble at this stage might be close to bursting it. In fact, this whole deal is in my opinion best described as ‘bonkers’. But what would I know?
You must be logged in to post a comment.