Since 2017, six companies have gone from c5% of worldwide market capitalisation to now just short of 20%. These are the well-known magnificent seven, ignoring Tesla. At the same time, they have gone from 4% of worldwide profits to an estimated 14% in the current year. An astonishing outcome and one that if you had predicted it would have been hugely profitable.

Earnings growth for these companies has been outstanding but so has the valuation expansion. The market is expecting for them to account for an ever-growing percentage into the future. Which implies that the rest of the world is expected to continue its decline as a portion of profits.

These companies are obviously hugely successful, they have become crucial to the world economy, providing core services for almost all consumers and businesses. In most cases, they are monopolies or at least dominant in their markets. There is a lot to like. From this position potentially, they might be able to continue to suck in profits from the rest of the world economy; an unregulated monopolistic utility position at the core of the world economy is a nice place to be. The market certainly agrees, particularly with the hype around the opportunity from AI going forward.

The chart shows the share of worldwide earnings from these mega cap tech companies since 2017.

Source: Bloomberg 29.12.2017 – 31.12.2024. Please note that the 2024 financial year end figure is estimated

There is however, a potentially negative side to these companies which isn’t perhaps, fully understood. These companies’ business models are changing. Capital expenditure as a percentage of operating cash flow has gone from 28% to 38% at Meta from 2017 to 2023 and from 20% to 38% at Microsoft. There is a similar trend elsewhere among these mega cap companies. This is because the ‘AI’ trend requires a huge amount of capital expenditure in data centres. They are no longer competing with product development and software; they see future success from massive capital investment in processing capacity. This capital cycle has only just begun.

This is a major change for these industries. AI needs to be a major revenue generator to justify the level of capital investment. Unfortunately, thus far there is little evidence of any viable revenue model that justifies the expenditure. Even if this arises, it appears these businesses have shifted to a capex heavy model, where success come less from innovation and more from the ability to spend more than your peers. If that is the case, in the long term, margins and returns will be greatly eroded over time, despite the huge market concentration. The historic level of exceptional returns may be in the past. Growth may continue but it is becoming a different type of growth, more akin to traditional industries, where incremental growth comes with incremental capex. Even concentrated markets can end up highly competitive in these circumstances.

The difference from the early years of the internet is clear- today’s big winners are largely involved in providing the tools used on the internet, not the infrastructure. The providers of the infrastructure were caught up in the TMT (Tech, media and telecoms) bubble, but ultimately proved unsuccessful. The current leaders are planning to provide the AI infrastructure, perhaps the ultimate winners will be those that lever these new tools to create new profitable business models.

We have always had an exposure to our digital economy theme, although portfolio construction discipline requires us to not have an excessive weight and to continuously trim as they rise. The momentum aspect of our process leads us to continue holding on and only should that momentum fall away would we sell. In the meantime, while all attention is on these companies making them difficult to ignore, we would emphasise that there are widespread opportunities worldwide in equities, which are also doing well.