Market Musings 261024: How To Invest In AI In 3 Indirect Ways

26th Oct '24 by Edmund Shing

4 comments 1656 reads

Market Musings 261024:
How to invest in AI in 3 indirect ways

Podcast: European carmakers on the skids (click on title for link)

The AI theme has been hot all year long

There is no doubt that the AI mega-theme has dominated 2024 stock markets, with the stellar performance of the Magnificent 7 (+47% year to date) closely followed by the US Semiconductor sector (+47%).

Magnificent 7 and Semiconductor sector both boosted by AI-related flows

However, I think it is prudent to raise a number of concerns for investors to ponder in relation to this Artificial Intelligence theme.

Firstly, how can we be sure that the huge investment being made by Microsoft, Google, Meta and Amazon (the Hyperscalers) in particular will generate high returns on investment, even in the long run?

History is littered with examples of disruptive technologies (think AI today) that are themselves disrupted by even newer technologies. The advent of the smartphone with Apple’s first iPhone launch in 2007 is such an example, resulting in the decline of then-leader Nokia’s mobile phone global market share declining from 44% to eventually zero.

Nokia’s post-2007 inexorable decline in mobile phone sales and profits

AD_4nXcboAuaO-IO23EYrEFZE6QBRWk7EuolxnZxRpdKUWtIyNEGDt6P0olIuYw-A8sfhsnfXeE4xSfbanNO1DU1KSRhEHntLGTxcXBB_4ZQYb3glSxUSY3ETnEvRJfdpQ9AGs1CjcfSwelnpXZBUklwKgWYKz0E?key=CmhGYNEpyH6RLM4AcdOluQ

Source: disruptiveinnovation.se

It is thus dangerous for investors to invest heavily in an emerging megatrend such as AI, given the lightning speed of evolution of the state of the art in AI and the enormous size of investment being poured into building new AI-related data centres.

AD_4nXcwg8mT0oKKnbT_Q57x_MPLIUMs-qkzk9S3nQXyl5xLvTKldAwMImX38GFZ95gWsh8VfhLfaN0-gNezvt9OkHEm6Dvt6zlLfJjh4FnjS4oUWkM55Q9YBTruvs8Jb4hI3HpohoVgQDbuubZUSprSr48UGTkd?key=CmhGYNEpyH6RLM4AcdOluQ

The risk of oversupply

Should the public’s interest in AI fall off, you can be assured that this spending level will fall, and there may even be an over-correction. This is similar to the technology boom-bust cycles that we have seen in the past linked to over-optimism and then over-investment, for instance in the year 2000 technology-media-telecom stock market bubble that saw huge over-investment in fibre optic networks, given inflated predictions of fibre-optic demand from surging internet traffic.

In actual fact, the predictions of surging demand for fibre-optic network capacity proved correct, but required the advent of audio and video streaming (think Spotify, Amazon Prime and Netflix) many years later.…

Unlock with Email

Unlock with Google

Already have an account?
Login here

snowdrop

26th Oct '24

Adam Toose has noted that projected demand for electricity is driven to a greater extent by a/ c and electric cars than data centres. A/c is hard to predict so you could add gas generation suppliers to your list. The term propriatory d/b covers the down stream potential of AI for firms to apply using their own software. I note you make few suggestions there. AI will eventually be driven by software but no one knows who the winners will be.

Howard Adams

27th Oct '24

Hi Edmund

I agree with your write up and have been building up some positions in what I term 'the AI stack', then the 'AI data centre and nuclear stack'

In the AI stack I hold the hyperscalers but I also focus on the top App. layer. That is the companies creating software products which actually make use of the outputs of LLMs in order to drive business decisions and operations. Clearly hyperscalers are moving towards this offering, but are quite behind in this layer of innovation, in contrast to my main holding Palantir Technologies (NYQ:PLTR) .

In my view Palantir Technologies (NYQ:PLTR) is an investment which requires in depth research in order to appreciate its potential, but if the concept of the 'Operating System' for the modern enterprise is an area of interest for someone, I think digging into Palantir Technologies (NYQ:PLTR) will be quite rewarding. Some believe it will replace Microsoft (NSQ:MSFT) within enterprises. We shall see.

With regard to AI data centres and Nuclear. I have starter positions in Applied Digital (NSQ:APLD) , Oklo (NYQ:OKLO) , Nano Nuclear Energy (NAQ:NNE) BWX Technologies (NYQ:BWXT) Rolls-Royce Holdings (LON:RR.) and several others associated with operating the infrastructure of these data centres.

Regards

Howard

For anyone interested in the whole AI, chip, data centres space this is an excellent presentation of the whole space. The presenter is the author of the quite recent book 'Chip wars' (an excellent read).

&t=1005s

Dragon

28th Oct '24

Cordiant Digital Infrastructure (LON:CORD) Is a trust focused on

"Digital infrastructure is the unglamorous, but essential, ‘plumbing of the internet’. These mobile towers, data centres, fibre-optic networks and wireless sensor networks undergird the wider digital economy."

All needed by AI.

Market Musings 261024: How to invest in AI in 3 indirect ways