Investing amid era-defining tech change and an expensive U.S. dollar

Views from the TIAA General Account

How does the recent rate decision by the U.S. Federal Reserve chime with the General Account’s market outlook?

The Fed’s 25 basis point rate cut on 17 September was widely expected and consistent with our outlook. The Fed signaled a bias toward reducing rates in the face of weak employment data and inflation holding in its current range. Our view is that the Fed will deliver one or two additional rate cuts by year end and perhaps further easing into the new year. The first half of 2025 delivered a material slowdown in economic growth relative to 2024’s 2.8% expansion, however, our central case does not see the U.S. slipping into recession.

From a broader markets perspective, both bonds and equities appear caught up in the euphoria of the moment, with the promise of fundamental technological advancements, improving productivity and spending on artificial intelligence (AI) all providing strong tailwinds.

We anticipate the yield curve will continue to steepen as the Fed rate cut campaign progresses. We do not see credit spreads moving materially wider with borrowers still seeing top-line growth, stable to improving margins and benefitting from lower debt costs. Both these trends would be good for long-term, book yield-oriented investors like the General Account (GA).

On balance, we have a constructive market outlook. But as long-term investors, we are vigilant about the potential for derailment caused by unsustainable fiscal deficits, growing geopolitical instability and tariff-driven inflationary pressures that are likely to build over time.

Many investors are re-thinking exposure to U.S. assets and the dollar. How does the General Account view its geographic exposure? What changes, if any, have been made?

Despite the U.S. dollar’s depreciation this year, it remains historically expensive sitting in the 92nd percentile of valuations over the last 35 years. Approximately 15% of the GA’s invested assets are outside the U.S. We manage a well-diversified portfolio with the preponderance of exposure within developed countries around the globe. Our exposure to emerging markets is moderate, with much of it in sovereign risk and in actively traded strategies where we expect to generate attractive risk-adjusted returns.

We feel we are well positioned, and we work closely with our Country and Macro Risk Team to ensure we are agile in adjusting and trimming unwanted risks as they crop up. We have been diligent in running what-if scenarios and developing potential pivots, given the large number of geopolitical headwinds this year. In terms of currency, we look to hedge nearly all our foreign currency risk back to U.S. dollar, which is consistent with our liabilities, which are all denominated in dollars.

With AI adoption ramping up, the demand for energy appears ahead of supply. Where does the GA see opportunities in energy infrastructure?

AI has the potential to be the defining technology of our time; automating routine work, augmenting human decision-making and accelerating breakthroughs that were once unimaginable – and, if imagined, way beyond our reach.

It is not simply a tool for incremental efficiency but a platform for transformation that will leave no industry untouched. However, its ultimate impact will depend on how responsibly it is developed, governed and integrated.

A constraint exists in its demand for power. For example, training a single large model can consume as much electricity as nearly 100 U.S. households use in a year.¹ Furthermore, serving AI applications at scale may drive more sustained energy demand than training. Projections I have seen suggest that AI could nearly triple data center electricity use within the next 10 years. All of this is reshaping utility planning (utilities are now operating with nearly no reserve capacity in the system) and challenging transmission infrastructure, which is quite dated in the U.S.

Clearly the need for infrastructure spending creates an investment opportunity, as does the need to support alternative forms of delivering power directly to data centers. As an example, I’d point to the growing need for liquid cooling technology rather than air cooling for the massive scale of processing units that generate much more heat per rack than the typical central processing units within these data centers. Hyperscalers, such as Google, Microsoft and Nvidia who have large-scale data centers providing extensive cloud computing and data solutions, are investing heavily in liquid cooling to enable next-gen AI models without hitting thermal limits.

As signs of a real estate recovery take hold, where does the GA see the opportunities and risks?

Sentiment that real estate values have found a floor in this cycle is gaining traction among investors, and global real estate returns have been positive for the past five consecutive quarters.² Over the trailing year, transaction volumes are also rebounding, with Q2 investments up 17% relative to the same period last year across the U.S., Europe and APAC.³ We are seeing a rebound in allocations to high-quality industrial and housing, keeping up with rising demand for new originations.

Given the earlier topic of AI, well-situated and well-equipped data centers look interesting. We are, however, mindful of the potential for grid bottlenecks, cooling requirements and the fact that a few hyperscalers dominate demand. Other sectors within real estate that appear compelling include grocery-anchored and neighborhood retail, self-storage and medical offices.