Hi everyone,
A quick question: out of the Mag 7, which do you think have the largest and lowest downside potential in case AI does not turn out to be the goldmine many think/thought it could be?
There are increasing concerns around the current version of AI (LLMs) being limited and not able to reach AGI (see, for instance, Yann Lecun’s take on it), coupled with concerns around monetisation (or, at least, monetisation in a scale compatible with providing good returns on the huge capital being invested), hence my question.
I think Nvidia is the top loser in that scenario, and Tesla the least exposed. Not sure about the other ones.
Thank you!
P.S.: My current Mag7 top pick for the medium-long term is Alphabet.
Mag 7: exposure to AI downside
byu/Puzzleheaded_Bat3349 instocks
Posted by Puzzleheaded_Bat3349
10 Comments
Google
AI could deliver low ROIC for one company and high ROIC for another company. It’s not all or nothing as some make it out to be.
I do believe tesla vill continue to struggle in either case with their trash fundamentals.
NVIDIA because if chip demands dies they will plummet.
Imo it’s not Tesla, it’s Apple. Apple will be the winner because they’ve spent the least and they can tell the market. We told you so and it wasn’t wise for us to burn our money like these other companies.
If ai is truly a bubble then i would say meta, google, Microsoft and Amazon are the worst off. They have invested the most in it. Nvidia makes hardware. Their risk is someone making better hardware. If ai is a bubble then all that hardware these companies bought is worthless. Nvidia can pivot to whatever people/companies want to buy. Nvidia has a backlog on hardware. Meta/google/MSFT/AMZN all don’t have back logs. They need to generate income from that hardware to justify it’s expense.
That all being said. I don’t think ai is a bubble and these are the best companies to benefit for the next decade. Leading with Nvidia.
It’s not mag7 but Oracle has taken on a lot of exposure to the downside risk by ordering GPUs and building AI datacenters. If the demand for their AI super clusters doesn’t materialize over the next 5 years they are going to be in a world of hurt. The stock has already priced in 10x revenue growth.
You don’t need agi to create incredible solutions with LLMs
Nvidias valuation can literally bounce between hundreds of billions on a weekly basis.
It’s NVDA.
I think Apple comes out ahead here. Competitors are pouring trillions into building compute capacity – given the mania, I imagine we’ll be looking at an eventual oversupply. The same way we overbuild fiber optic cables during the dotcom bubble.
Meanwhile, Apple is effectively letting others subsidize the cost of generative compute. Once these massive data centers are built and compute becomes cheap and abundant, Apple can move in and seamlessly integrate it into their hardware—especially within visionOS.
u think too much
I think the run on AI will continue maybe one, two years, then it will implode. Nvidia makes the most margin of all, they might fall the hardest. Because the stock price is a bet on the margin and guidance. But it’s capped… they cant ramp up production forever, the investors cant pump money without end and finally the energy cap.
You know Gemini and Copolit and many others are for free and when I aks Gemini to draw me a pic it consumes electricity. Half a kWh, it consumes time shared software and hardware costs but all of that is offered for free to a billion of users on earth becuae the big techs believe they have to do so, to monetarize that later.
Will there ever be a return? Noone can say. Firms like Palantir, which do really create value with AI or Duolingui, ServiceNow where AI is replacing lot of human workforce – yes they do. Even if they have to pay $1000 a month for a particular AI model…. it’s cheaper than a $5K white collar worker.