Why long-term US rates are rising

When the yield on the 10-year US Treasury rises, the first question is often what the Federal Reserve will do. Robert M. Almeida, portfolio manager and global strategist at MFS Investment Management, proposes another: how much capital does the economy need today to build the productive capacity of the coming years? In his October 2026 note, the author does not deny the weight of inflation or monetary policy, especially in the short term. But he argues that, to interpret long-term rates, one must also look at demand for capital. If many companies and institutions must finance projects at the same time that absorb resources for years, the price at which that capital is available can change even without a new Fed decision.

The Fed and the 10-year Treasury answer different questions

The Federal Reserve steers the cost of money at the very short term and, through its decisions and communications, influences financial conditions. The yield on 10-year Treasuries, however, is formed in the market. Anyone lending money for a decade takes into account expected growth and inflation, the possible path of short-term rates and the remuneration it judges appropriate for taking the risk of holding a long-dated security. In Almeida's reading, the balance between saving and investment also matters. This distinction explains why a change in the official rate does not necessarily translate into an equivalent move in the 10-year: the two maturities reflect, in part, different questions and risks.

Why the 2010s needed less capital

In the 2010s, observes Almeida, many companies could boost profits without expanding their physical capacity proportionately. Outsourcing production and global supply chains made it possible to use plants and skills already available elsewhere. In some sectors, software and intellectual property mattered more and more in the creation of value, while building factories mattered relatively less. Some of the resources not needed to finance new projects could be returned to shareholders. According to MFS, demand for funding, contained relative to globally available savings, thus helped keep real rates and long-term bond yields low. Central banks reinforced that backdrop, but in the author's argument they were not the only force determining it.

Today many investments compete for the same resources

The backdrop described in the note has been turned on its head. Data centres and semiconductor factories need financing, but so do power-generation plants, electric grids, defence capacity, manufacturing sites and more resilient supply chains. The point is not just the size of individual investments. It is that many start in the same term and compete for similar resources: capital, energy, labour, equipment and time. They are also strategic projects that often cannot be scaled back quickly once under way. If available saving remains plentiful but grows more slowly than demand for capital, those financing a project may be able to demand more generous conditions. The chart in the note illustrates this shift in the balance; it does not show how many yield points are attributable to each factor.

Artificial intelligence is absorbing capital upstream

Artificial intelligence makes the shift to an economy that requires more infrastructure plain to see. The user sees a digital service, but to deliver it there are data centres, chips, memory, power supply, cooling systems and grids. These structures require investment before the service generates revenue and can tie up companies for years. In Almeida's reading, then, the growth in demand for artificial intelligence does not imply only more spending on software: it entails demand for physical capital. The distinction matters when reading balance sheets too. Increasing productive capacity can create value, but the result depends on how much that capacity will be used and what revenue it can support relative to the capital employed.

Downstream, competition rises and expected returns change

The other half of the argument concerns those using the technology. Artificial intelligence can reduce barriers to entry and allow new competitors to reach activities that once offered relatively protected profits. An established company may then decide to invest to win a new market, or be forced to do so merely to defend customers, market share and margins. In both cases the spending may look similar on the balance sheet, but its meaning is different: in the first it supports expansion, in the second it may serve to limit deterioration. The paradox identified by Almeida is that companies may need more capital just as competition makes the return on that capital more uncertain. Growth in investment, then, does not by itself guarantee corresponding growth in profits.

What this means for investors

From this comes the distinction the note offers investors: separate companies capable of earning attractive returns on new investments from those increasing spending chiefly to preserve margins. It is a distinction that can change market leadership. The companies rewarded in the previous decade often managed to grow with little additional capital and to preserve their sources of profit for a long time. If both of those conditions become less common, yesterday's winners are not necessarily tomorrow's. Almeida sees this greater differentiation as an opening for active management: comparing the quality and return on investments rather than treating all companies exposed to the same technology as equivalent. It is an opportunity for selection, not a guarantee of results.

How to read the thesis without turning it into a forecast

The note's contribution is to explain why the cost of capital deserves attention even when the debate centres on the Fed's next meeting. It does not follow that long-term rates must rise without interruption: growth and inflation expectations, financial conditions and the availability of saving can change and push the yield in different directions. The useful rule for market readers is to keep two questions together. The first is how much it costs to finance new investments; the second is how much those investments can earn, taking competition into account. It is the link between the two, in Almeida's reading, that connects Treasuries, company valuations and changes in market leadership.

Source: Robert M. Almeida, MFS Investment Management, Strategist's Corner, October 2026 — Why US Rates Are Rising