At NEPC, we believe asset allocation is integral to the success of any investment portfolio, and it is among the most important decisions an investor will make. In a year that has been full of surprises so far, we offer you a behind-the-scenes glimpse of the issues and themes at the forefront of the meetings of the NEPC Asset Allocation Team.
The participants:
Jennifer Appel, CFA, Principal, Senior Investment Director
Robert Goldthorpe, ASA, Senior Investment Director
Christopher Hudson, Senior Investment Associate
Phillip Nelson, CFA, Partner, Director of Asset Allocation
James Reichert, CFA, Partner, Senior Director of Portfolio Strategy
Michael Zampitella, CMT, Investment Director, OCIO
Aparajita Bubna, Managing Editor, facilitated this roundtable
1. What is your view on the U.S. midterms’ election results?
Phillip: I think the federal election results were generally consistent with what the polls had been suggesting, although the magnitude of the shift across state legislative and governor races was much larger than expected.
War fatigue and gas prices were clearly significant factors, but the political swings in data center hot spots such as Texas, Georgia, and Ohio reinforce how intensely many voters feel about the buildout off data centers in their communities.
The 2028 election cycle effectively begins now, and any legislative progress is likely very limited over the next two years with both parties looking to the future presidential race and the Democrats holding dreams of a filibuster proof 60 seats in the Senate. While divided government is generally a favorable result for U.S. equity markets, markets will begin to discount the range of 2028 outcomes which could bring additional volatility.
2. Data centers were a flashpoint this election cycle. What do you make of the election results?
Jennifer: There may be heightened focus around specific legislation or regulations that could affect data-center development. Governors, state legislatures, and local officials have some ability to pause or delay projects, but we do not see the infrastructure buildout for AI ceasing but maybe just slowing a bit.
That said, I would reiterate a point we often make about elections: we generally don’t view them as having as much influence on markets as broader macroeconomic factors such as growth, inflation and monetary policy. Those are the forces that tend to drive markets over the long term. As long as the current economic catalysts remain in place—including a resilient consumer, healthy labor market, and continued AI investments—I don’t think the election outcome is likely to derail the broader economy. Ultimately, I think the general view is that gridlock tends to be positive for markets.
3. Is the AI boom creating a sustainable earnings cycle or a valuation bubble? What could trigger a meaningful market correction?
Michael: Like any market theme, there are winners and losers, so certainly both. Large-cap technology stocks are under pressure, but forward P/E multiples looks somewhat attractive today. Data centers buildouts are still in the early stages, and the returns remain somewhat unproven. It’s a favorable environment for active management.
Jennifer: From an earnings perspective, certain areas of the AI market are unlikely to maintain their current levels of growth over the long term. Memory is a good example: compelling economics attract new competitors, which may moderate earnings growth over time. This does not mean earnings growth will disappear, but the current pace is unlikely to persist indefinitely. Other areas of the AI buildout may be more protected by physical constraints. As a result, we expect earnings opportunities to become more selective moving forward.
Phillip: The question of whether AI is creating a sustainable earnings cycle is particularly interesting. While the current concentration of spending among hyperscalers is unlikely to be sustainable indefinitely, I believe we are very early in the AI buildout. For these investments to generate sustainable returns, an entire ecosystem and economic structure must develop around the AI buildout. In the same way that when you build an electricity grid, the value extends beyond simply delivering electricity from one point to another. The grid paves the way for productivity gains and supports the development of an ecosystem of businesses and applications that build on that infrastructure. Despite the enormous amount of capital already being invested, we may be only in the first or second inning of this cycle.
James: Thinking about what could trigger a meaningful market correction…serious competition challenging the largest AI names and pressuring profit margins is a risk, but the competitive moat of many companies may make that difficult. A meaningful market correction is hard to imagine if companies continue to generate strong cash flows over the next several years. Regulatory intervention could slow the growth potential, but neither higher interest rates nor geopolitical risks seem likely to derail the AI boom. Ultimately, a meaningful correction may require a slowdown in the actual cash flows and capital expenditures underpinning market valuations.
Phillip: I think the regulatory outlook is probably the biggest overhang. It may not necessarily pose a fundamental risk to the AI theme, but it could slow the pace of investment and growth.
Jennifer: I agree, particularly given the midterms, data centers were an issue. For local politicians looking for an easy win, taking a negative stance on data center development seemed to be the way to go. That creates a potential risk, as James mentioned, by slowing development and increasing the lag time of building out the AI infrastructure.
James: Opposition to data centers may be one of the few bipartisan positions. That said, some communities are already being offered financial incentives to host data centers, and the longer-term implications of that approach remain to be seen. Onsite energy production could also emerge as a potential solution. While greater political scrutiny may weigh on margins in the near term, they could ultimately lead to a better model for the technology and its longer-term productivity benefits.
Phillip: Another potential catalyst for a market correction is increased competition, whether from competition in the GPU chip market or through a significant decline in memory prices. Lower costs could enable more flexible pricing, but also put pressure on profit margins and valuations. While that could create near-term headwinds for some stocks, it would likely be a healthy development for the broader AI markets and, ultimately, AI productivity gains over the longer term.
4. When you look at forward market pricing for AI-related stocks today, many of the technology names do not command a market premium. What do you think is holding them back?
Jennifer: Given the fundamentals of many companies central to the AI story, you might expect them to command a higher valuation premium, but that’s not what we’re seeing. This likely reflects the broader lack of market enthusiasm and negative sentiment. Current forward valuations suggest investors remain somewhat skeptical that AI’s benefits will translate into the longer-term earnings growth many are anticipating.
Phillip: Is the skepticism really about whether AI will deliver the expected cash flows and productivity gains, or is it more of a generalized fear of AI? There’s almost a sense that AI is the boogeyman, driven by concerns about job displacement and an almost irrational fear of what this technology could ultimately become.
Christopher: From a productivity standpoint, an important consideration is whether the gains are coming primarily from a declining labor share. If AI-driven productivity gains come at the expense of labor, that could put pressure on consumer demand over time.
James: AI investment may have helped sustain economic growth and offset some of the drag from higher interest rates over the past few years. Historically, consumption has been the primary engine of the U.S. economy, accounting for roughly 70% of GDP, but we’ve become increasingly reliant on technology and AI investment to drive growth. If that spending eventually moderates, whether due to limits on growth or regulation or other factors, more of the burden will shift back to consumer demand, creating a potential vulnerability in the medium term. If AI-driven job displacement isn’t offset by new business formation and job creation, the AI transition could weigh on household incomes and consumption.
Robert: I would zag and say if AI delivers meaningful productivity gains, particularly across corporate America, the outcome may be a shorter workweek rather than widespread job displacement. We’re already seeing a gradual shift toward shorter workweeks, and that trend could eventually lead to a three-day workweek. Humans will still be needed for executive decision-making, oversight and governance. In fact, if productivity gains translate into fewer hours rather than fewer jobs, we could ultimately see labor shortages rather than job losses.
I think the more important question is what do the balance sheets look like as companies scale AI? Are businesses becoming more asset-intensive, requiring significantly more capital to sustain growth? Even if earnings continue to grow, the capital required to scale that growth matters. To me, the market’s longer-term concerns are more about the underlying economics, and balance sheet implications.
Phillip: That’s spot on! It really comes down to a simple question: three years from now, will the return on invested capital for hyperscalers and other major AI spenders exceed their internal cost of capital? If AI is genuinely enhancing productivity, the returns should justify continued investment. We’re already seeing early signs that return on equity remains high across many technology companies, suggesting the spending cycle has room to continue. In fact, I would argue that we’re still not seeing enough investment today relative to the opportunity ahead.
James: I will challenge your idea of a shorter workweek! A three-and-a-half-day workweek assumes labor has gained enough power to capture productivity gains in the form of more leisure time. But in a competitive environment, those gains could just as easily be competed away. Companies that maintain a five-day workweek may simply outperform those that reduce hours, particularly when there are always competitors willing to work harder and capture market share.
For a shorter workweek to become sustainable across corporate America, productivity gains would likely need to translate into a much broader increase in economic surplus and quality of life. Otherwise, I think competitive pressures would make that outcome difficult to sustain. There’s always going to be someone hungry enough to eat your lunch.
5. How vulnerable are diversified portfolios to continued concentration in mega-cap technology stocks?
Jennifer: I think the key point about market concentration is that this isn’t just a public equity issue. There’s a broader macroeconomic dimension, with AI investment influencing interest rate expectations, growth projections and overall market sentiment. We’re also seeing increased exposure in investment-grade credit. So, diversifying away from AI isn’t as straightforward as simply reducing exposure to a handful of large-cap technology stocks. AI is increasingly embedded across asset classes, including private markets, making it a much broader portfolio-wide consideration.
Phillip: There is a specific index-concentration concern: if the top 10 companies grow to represent more than 50% of the S&P 500, that may be a sign of an unhealthy market structure, and we would advocate taking steps to diversify S&P 500 exposure.
But when it comes to broader vulnerability to AI, virtually every portfolio has exposure in some form. AI influences energy markets, private and public markets, inflation, real interest rates, economic growth and Federal Reserve policy. There really is no escaping it, which makes identifying a straightforward or even practical hedge against AI extremely difficult.
The investment landscape has fundamentally changed. Investors need to adapt by finding thoughtful ways to participate in the opportunity while recognizing that markets could look significantly different over the next decade.
Robert: Many of the largest technology companies are diversified businesses. Owning them already provides exposure to a range of technology markets. The risk is whether they can manage those different businesses effectively enough to continue generating strong returns on investment, particularly as they grow larger.
Phillip: I think it’s fair to say that we may be too focused on current market dynamics. Ten years from now, the top 10 companies in the S&P 500 will likely look very different; there may be an entirely new group of companies in that mix that we haven’t heard of yet. In the year 2000, the top five names of the S&P 500 were Microsoft, General Electric, Cisco, Exxon Mobil and Intel. Now we have Nvidia, Apple, Microsoft, Amazon and Google.
James: It is hard not to mention though that those five stocks make up more than 25% of the largest index of the world. That’s something!
Phillip: I think anyone in the investment industry would be foolish to disregard the quarterly forecasts and earnings calls of the five largest companies in the S&P 500. Their economic impact is comparable to that of many countries, if not greater. They are macroeconomic forces in their own right, and we shouldn’t dismiss them as fad-bubbles simply because they are relatively new or have grown so large.
James: If you ranked the S&P 500 by number of employees rather than market capitalization, I think the top 10 companies would be just as economically significant. That also connects to the earlier points about the consumer and how productivity gains may eventually flow through the economy as large employers begin to reduce headcounts. But I suspect that process will unfold gradually, rather than through the kind of widespread layoffs that could trigger a recession.
Phillip: I think with any new technology, we tend to focus first on the negative, particularly the jobs that will be lost. But what about the new economic activity and jobs that the new technology will create? Those opportunities are harder to envision because they lie in the future, so you have to have some faith that they will emerge. But if the technology develops into a sustainable economic enterprise, it should create entirely new industries and avenues for people to work.
6. Where are the most compelling opportunities beyond AI leaders, and how should investors balance participation in the upside downside protection?
Phillip: Given how the market is priced today, I think investors are overly focused on finding downside protection in their equity exposure and AI-related investments. I would say, let them run and let them grow. I see a great deal of hesitancy from investors today. If you look beyond the AI landscape, equal-weighted S&P 500 earnings growth has seen double-digit growth over the past 12 months and is expected to continue over the next 12 months. The broader economy is growing, and that growth isn’t being driven by just a select number of names.
James: Small- and mid-cap stocks have been under pressure recently, in part due to higher interest rates. To the extent the upward pressure on interest rates moderates, U.S. small- and mid-cap stocks could present attractive opportunities to capitalize on growth over the next couple of years.
One other point: Phill mentioned letting the AI names run. Many of these companies, as we have discussed, are among the largest and most liquid in the world, which also makes them relatively easy to rebalance. As your equity exposure approaches or breaches the upper end of your target range, taking some profits and rebalancing into the safety of Treasuries is a reasonable approach, particularly with yields as attractive as they are today. But I wouldn’t be underweight equities; continue to maintain an overweight position relative to your targets, while having risk controls in place to ensure you rein things back when exposure gets a little too high.
7. Could the Iran-U.S. conflict trigger a global energy crisis if supply disruptions persist and strategic reserves are depleted? How could a sustained oil and gas shortage affect inflation, global growth and central bank policy, and which asset classes would be most vulnerable – or best positioned to withstand the resulting market shock?
Jennifer: I do think this is an underappreciated risk in markets today. For much of this year, markets have been waiting for a dislocation in oil prices, potentially reaching $200 per barrel or more. But if you look back over the last few months, there have been a number of buffers in place in the economy that have helped keep energy markets in check. As we get toward the end of the year, many of those buffers are starting to run thin, whether it’s global reserves or increased pressure on energy infrastructure in the Middle East. Now, China is starting to come back to the market online, even though it has been managing demand and draining energy reserves, which has helped keep prices lower. I think the natural inclination is to view the risk as highly inflationary for markets, but I would push back on that somewhat. At a certain point, demand destruction becomes part of the story.
8. What is your view on emerging markets?
Jennifer: This is not your father’s emerging markets portfolio. It’s very different from what many investors originally underwrote when they added it to their portfolios.
There is now significant AI exposure, with Taiwan and Korea representing more than 50% of the index. So, you’re dealing with different return drivers and, potentially, a different diversification profile than in the past. That said, I still think it comes back to the importance of active management, particularly given the concentration that now exists within the index beta.
Michael: If you’re looking to diversify away from the Magnificent 7, you should probably steer clear of emerging markets. Three names alone make up 27% of the index. So, to Jenn’s point, active management is important. The breadth of the market has been abysmal, which could also create an opportunity for alpha.
James: I think it’s much the same as in the U.S., right? The concentration at the top makes it difficult for active management to succeed unless some of those top names underperform.
An interesting area we’re looking at in both markets is active extension strategies. These give managers the ability to take both long and short positions, potentially helping them identify mid- and small-cap names within the AI theme that may be over- or under-valued, while complementing exposure to the larger market-cap names.
These extension strategies offer investors another way to maintain an active approach in highly concentrated markets, particularly when we’ve seen some of the more traditional fundamental strategies struggle to outperform over the past couple of years because of that concentration.
9. What are the key forces likely to shape markets in 2027?
Robert: Interest rates. Are there more hikes to come, or are we looking at rate cuts? Will the market accept Treasury yields at current levels, given the pace of government spending?
There’s a lot going on here. Some of these dynamics have already started to bubble up, and I expect them to remain a key part of the story in 2027.
Phillip: I think an issue for 2027 is the extent to which high interest rates pose a headwind to private markets and the ability for companies to find exits and liquidity. We are several years into a cycle where distributions have been slower than investor expectations, and that stress remains.
Michael: I’ll give you three: AI monetization, the political landscape, and the Fed.
Jennifer: I also think the geopolitical backdrop matters for 2027. We were talking about the potential for disruption in energy markets, and the impact on inflation and policy.
Chris: My main force would be Fed policy and any volatility that comes from that.
10. Looking ahead to 2027, what is the single biggest risk that you believe markets are mispricing today?
James: I think the single biggest mispriced risk is the concern that the Fed may not remain independent over the long run and won’t provide the appropriate support for markets. Right now, that independence is being questioned to some extent, which is putting upward pressure on rates. I think there’s too much concern around this uncertainty about the Fed’s long-term role, and it’s pushed rates too high.
I’m not necessarily making duration my number one recommendation, but I would stay long equities because they should benefit from a tailwind of lower rates as some of these concerns get priced out. I also wouldn’t be overly concerned about the growth trajectory because these higher rates could ultimately prove fleeting.
Phillip: Treasury yields are higher not because inflation expectations are rising, but because real interest rates are higher. The question is how much of that reflects strong economic growth in the U.S. versus a higher risk premium that some investors are demanding.
I think investors should continue to view Treasuries as a safe-haven asset. Ultimately, I expect rates to come down somewhat, which would benefit U.S. small-cap stocks and other interest-rate-sensitive sectors.
In some ways, I think the continued growth potential of U.S. corporations is the most mispriced aspect of markets today. There’s too much doubt and concern about the growth outlook, not just in AI, but across many areas of the U.S. economy.
Robert: I thought the biggest risk heading into 2026 was rates not coming down. But rates climbed higher, so new deals have to be financed at today’s rates. Are investors still assuming rates will come down over the near to medium term? If the NPVs no longer make sense at today’s yields, does that slow private markets? Does it slow the AI buildout? I don’t think it necessarily slows AI, because the growth rates are high enough that the economics may still work in that space.
Are rates being held higher, in part, because of the enormous financing needs associated with the AI buildout? And is that potentially crowding out investment in other sectors of the economy that simply can’t support the same growth rates or justify financing at today’s costs? I think that’s an important risk to consider.
James: I think the AI deals that don’t necessarily make sense are the ones that were overleveraged and don’t have strong customers on the other side. If that means some of the weaker data centers don’t get built or completed, that’s okay.
Also, I don’t think everyone is assuming these deals will operate on razor-thin margins. I think they’re financing them with debt because they expect to benefit from a growth multiple. It’s not necessarily true that interest rates staying high impairs the economics of AI.
11. What is your most important investment recommendation to investors?
Michael: My most important recommendation to investors, honestly, is to ignore the headlines, because there’s a lot out there. Ultimately, we try to assign reason to a lot of market movement where there really isn’t any. So, sometimes it’s best to just ignore the headlines. Look for opportunities to rebalance, and ignore some of the narratives out there.
Phillip: To add to Mike’s point, sometimes the simple answer is the right one. Traditional core fixed income looks attractive today, and U.S. equities are reasonably valued. Both the S&P 500 and the Bloomberg Aggregate Bond Index are more attractive than they’ve been in many years. Sometimes, the “easy button” is a good answer.
12. What is your top performing asset class for 2027?
James: Long credit.
Jennifer: U.S. small-cap stocks.
Michael: U.S. technology stocks.
Robert: Cryptocurrency and I’ll use Bitcoin as a proxy.
Phillip: NASDAQ.
Christopher: U.S. equities.
Phillip: But as we have seen in 2026, markets can surprise us with episodic peaks of volatility and that’s probably the easiest prediction for what we could see in 2027.



