The Cost of Capital

“If it’s good, it’s all good. We’re great. But if it’s bad, it’s really, really, really bad.”

That was Meta CFO Susan Li’s warning about the company’s AI investment, as recalled by Chief AI Officer Alexandr Wang (29) on a recent episode of the Cheeky Pint podcast. Mark Zuckerberg described the commitment just as plainly: “We’re going to write this check for like, I don’t know, like $100 billion, $200 billion.” His question to Wang: “Alex, are you sure this is going to work?”

The exchange was humorous, but the financial stakes were clear. Enormous sums are being committed before returns are known. Even those making the decisions recognize the distance between the opportunity they see and the outcome they need.

That uncertainty extends well beyond Meta. The AI buildout is creating substantial capital demands at a time when governments worldwide also require ongoing financing. Both need investors willing to supply the money, and the price those investors demand matters to everyone else.

Let’s go back and relive the 2010s. Capital was abundant and cheap. Many of the largest technology companies were generating substantial cash flows from businesses built on earlier investments. Interest rates remained unusually low for much of the period following the financial crisis, supporting the economy while also encouraging borrowing and higher asset valuations. Inexpensive refinancing became an assumption embedded in financial decisions.

Today, those companies are committing more of their resources to an expensive new undertaking. Unlike software that can be distributed cheaply once developed, the infrastructure behind AI requires data centers, advanced chips, power generation, and transmission lines. These involve substantial spending, long lead times, and resources that other parts of the economy also need.

Some of the companies leading this investment can finance much of it themselves. But the buildout extends to utilities, equipment suppliers, and developers with very different financial resources. As more projects seek outside financing, the competition for capital could become more consequential.

The anticipated rewards are large enough that some borrowers may accept higher financing costs to stay in the race. Governments accustomed to cheap money have less flexibility: they must refinance existing debt, and persistent deficits require additional borrowing.

That is a lot of capital the world needs to find.

AI may ultimately produce meaningful productivity gains, expanding economic capacity and reducing costs. We see considerable opportunity in that possibility. The challenge is that much of the spending must happen before those benefits arrive.

In the meantime, greater competition for financing could keep borrowing costs elevated. The Fed controls its short-term policy rate, but longer-term financing costs also reflect inflation expectations, the supply of debt and the returns investors demand.

This brings us to another concern: policymakers may have less room to cushion the consequences if conditions deteriorate.

For decades, investors became accustomed to central banks cutting rates and governments increasing spending when economic or financial conditions weakened. That expectation became part of the investment environment.

A recent Deutsche Bank report, The End of the Safety Net, argues that those responses face greater constraints today. Persistent inflation complicates rate cuts. Higher debt burdens, ongoing deficits, and higher interest expense make additional government stimulus harder.

The tools remain available, but using them involves harder choices. Cutting rates to support growth could prolong inflation. Borrowing more to support the economy could push financing costs higher elsewhere. Investors should be careful not to assume the response to the next downturn will resemble the last one.

The Fed has already begun raising rates again in a new campaign. Its September 16 increase brought the target range to 3.75%–4.00%. The CPI reading available at that meeting was 3.4%. The first increase is behind us; the question is how far this campaign might run.

We reviewed earlier tightening campaigns to put that question into perspective, considering both inflation at the first increase and the interest rate already in place. On that simple comparison, today’s starting position resembles 1994: inflation was slightly below the policy rate when tightening began. The Fed ultimately raised rates by three percentage points.

Our sample is small, and the results depend on how we define the campaigns. That resemblance does not provide a dependable forecast. It does, however, give us reason to consider a 300-basis-point campaign a scenario worth preparing for.

Including September’s increase, that would mean another 275 basis points, bringing the target range to 6.50%–6.75%. We do not know whether inflation and economic conditions will require that much tightening. We can, however, consider what it would mean for the portfolio.

Higher financing costs can pressure earnings, make refinancing more difficult, and reduce the prices investors are willing to pay for future growth. If policy support is also more constrained, the adjustment could be less forgiving than investors expect.

There are constructive outcomes as well. Inflation could ease, productivity could improve, and the investment underway could expand the economy’s capacity. We want to participate in those opportunities while maintaining the ability to respond if conditions deteriorate.

With all of that in mind, we are continuing to move into a more defensive position. Approximately a week ago, we raised cash to 20% across all portfolios. Time will tell whether we need to move further.

That reserve provides some downside mitigation while providing capacity to invest when prices become more attractive. Holding more cash carries a cost if markets continue to rise, and we recognize that trade-off.

We do not need certainty about the eventual path of interest rates to make a measured adjustment today. Capital demands are growing, while the policy support investors have relied on faces greater constraints. We believe that combination calls for greater care in the risks we take.