Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
It is time to have the outside of the Edelsten family home painted. The first quotes are in and I am now encouraging my children to abandon their careers and move into the decorating business.
Keeping the exterior of the house in good shape is what you might call necessary capital expenditure — or “maintenance capex”. It is what we have to spend to keep its value but does not add to it. We did have the kitchen extended years ago and probably added half what it cost to the resale value. So half of that kitchen capex is “maintenance” and half “growth”.
Companies grapple with these kinds of capex challenges too. What is a good use of our money? What is a good use of company profits?
Now feels like the time to look particularly closely at company spending plans. It seems to me that an awful lot of companies need to spend an awful lot of money just to stop the metaphorical window frames from rotting.
Some are asking for government help. In the UK, the steel and water industries have been noisiest in claiming to be going bust. Medium-sized manufacturing companies in Germany also claim they cannot survive in the face of current energy and labour costs, as are technology companies in France.
It is easy to give these companies a wide berth. But a harder judgment call is required elsewhere. Long-term investors have their work cut out in assessing companies’ capex plans. The amounts invested have risen sharply in recent years, with large US technology companies ploughing hundreds of billions of dollars into the infrastructure for artificial intelligence, capex levels which make my painting bills look modest.
We want companies to invest for the future. We select stocks because we like their products — so if they make improved products and more of them, so much the better.
Furthermore, capital invested is not taxed (as it is if paid as dividends), so the returns on that capital invested can grow tax-free within the company.
However, sorting out how much of the capex is targeting growth and how much is for maintaining current assets is not straightforward.
For traditional manufacturing businesses with hard assets, the value of the plant is written down over its useful life by a portion of its costs. This is an expense — a charge called depreciation — that is deducted from revenue and so affects profits. In effect, the accountants are acknowledging that at some point this plant will need replacing and the company should factor some of the costs into the accounts each year.
But the implied assumption is that replacement costs remain pretty much the same. After the period of inflation between 2022 and 2024, the figures appearing in the books will underestimate replacement costs — so for many companies the profit numbers need a haircut as the depreciation numbers in the accounts are unrealistically low.
Second, some capex that appears to target future growth may actually just be aimed at keeping up with competitors. If all the widget-making companies invest in a new AI process to manage their inventory, their competitive environment will be unchanged, as will their margins. Their boards will, of course, believe they have invested in technology and modernised their business, but the best that could be said might be that they would be bust if they’d not spent this money.
In my experience, management teams generally exaggerate how much capex will lead to growth — not just in manufacturing companies but in technology companies. Although these have fewer tangible assets, they risk losing their audiences if they do not keep their software or media content fresh.
Unfortunately, Trump tariffs are likely to make future capex decisions less attractive to shareholders. Large US companies are making a range of commitments to move manufacturing onshore to avoid tariffs (and to give the President headlines). These new plants are bound to have a lower expected return on capital than offshore ones. Construction costs will be higher, as will the running costs if many workers are employed. Plus, you might consider this unnecessary duplication of effort. So shareholder returns will fall — and share valuations are likely to fall with them. These issues are most pressing for US-domiciled companies, whose shares are often on valuations that could fall further than most.
While tariffs are mainly applied to hard goods, returns on capital invested are also under pressure in technology stocks. Nerds like me note that Microsoft and Amazon both used to depreciate their data centre kit over four years. In 2022, both decided these data centres might last six years. In the case of Microsoft, that boosted their stated earnings in 2022 by a cool $1bn. In 2025, Amazon decided to write its costs down over five years. I admire the company for taking a more cautious view of how frequently it might have to replace the kit to keep up with the speed of innovation. It reduced the group’s stated earnings by $700mn. I’m not sure all the analysts who praised Microsoft earnings last week have checked the boost they get from believing their data centres will last longer than Amazon’s.
Microsoft is also becoming less a software company, with the extremely high return on capital of the Windows and Office businesses, and more of a data-processing company, which has much lower returns on capital investment. That said, we own a modest amount of both stocks, as both seem more able than most to cope with current challenges.
In short, then, look closely at capex. Do not assume it is fully factored into profit forecasts and be cynical about management boasts that this investment will make you richer. It is the company profits they are spending — profits that they could be giving to you as dividends, and money that might come in handy if you have your own capex needs.
Simon Edelsten is a fund manager at Goshawk Asset Management