Back in March 2025, when I was experimenting with ChatGPT and other AI systems, I wrote a longish article (Deep Dive – March 2025) attempting to anticipate, or at least explore, some of the implications of AI for investors, at a very general level.
Today’s financial news prompted me to recall that I wrote the following:
“The point here is that I have bypassed all of Google’s advertising! This hints at a significant devaluation in the shares of companies like Google and Meta whose main revenues are from advertising. One begins to appreciate why Microsoft has put so much money into this new way of using the internet.”
At the time, this was an observation about a possible consequence of AI rather than a fully developed investment thesis. But the underlying idea was that AI could do more than introduce a new technology: it would change the way people access the internet, undermining some of the economics of the existing advertising model.
Now consider the financial news of this week.
Google’s parent, Alphabet, has seen its share price tank after it reported slower-than-expected growth in search revenue while spending enormous sums on AI infrastructure. Investors have become increasingly concerned about the cash required to fund that investment and the possibility that AI is taking business away from conventional search. Microsoft, by contrast, has demonstrated strong growth in its Azure cloud business and generated substantial free cash flow despite its own heavy investment. Meta, meanwhile, whose advertising is not so dirrectly related to Interenet searches reports that its advertising business remains robust, but the cost of its AI ambitions is becoming an increasingly important consideration for investors.
The market is now grappling with precisely the sort of questions I was raising eighteen months earlier.
To be clear, I did not predict the timing of a share-price fall, identify its eventual scale or establish that Google and Meta were certain to be long-term losers. Nor did I foresee all the complications of the AI investment boom and I was very clear in the text that I was attempting a technical analysis of a major and rapidly evolving step-change in technology. I did not, in any way, work out that there would be a bubble in chip maker share prices resulting from data centre build out.
But I think the original observation deserves some credit. I deduced that if AI could answer questions directly, bypassing the conventional search process, the consequences for then current advertising markets did not look positive for Google.
I suggested that this was an issue that could reach into the economics and valuations of some of the world’s largest businesses. Incidentally, again taking note of news about job losses announced at Microsoft this week, I had also commented with some accuracy on the prospects for employment in certain businesses.
Of course, fund investors do not pick stocks and we can only hope that our fund managers also stay ahead when technology is changing rapidly. The well known technology fund manager Ben Rogoff once said at a conference I was attending that a good technology fund manager is always aware not just of what IS happening in the world of technology but what is ABOUT to happen!
