The investment question has changed with it. The first phase of AI investing was led by hyperscalers, advanced semiconductors, and large language model developers. The current phase has been driven by infrastructure buildout, and markets have already recognized much of that trade. Over the next few years, we expect the most attractive opportunities to center on more persistent bottlenecks, particularly around power, and on companies that control workflow, own the customer relationship, bring domain expertise, and turn AI output into business action. That includes the software infrastructure and applications that will shape how AI is deployed and managed. As companies integrate AI into workflows, adopters across industries should benefit. We expect much of the next phase of value creation to come from emerging AI-native companies and disruptive business models, though at this early stage, even today’s disruptors may be displaced.

In this edition of VantagePoint, we focus on three questions:

• Which bottlenecks are durable?
• Can rising revenues justify the capital required to sustain leadership?
• Where can value persist as AI becomes cheaper, more capable, and more widely available?

The broader consequences for labor and society may prove profound, but they remain harder to observe clearly and are progressing slower than the technology itself. Regulation and sovereignty risk considerations clearly matter, as does society’s willingness to absorb the pace of change and its consequences. Those forces will shape AI’s development, but they are not the focus of this paper. We focus instead on where the evidence is strongest today and where the investment implications are becoming harder to ignore.

Tech is still early, moving fast

AI capabilities continue to improve at breakneck speed. The leading labs remain in a tight race, and each major release raises the bar while increasing disruption risk for incumbents and start-ups alike. Leadership among US models continues to oscillate. Chinese models have narrowed the gap on several technical measures despite US export controls on advanced chips. The open-model ecosystem, much of it coming from China, has also become a credible lower-cost option for use cases outside of mission-critical workflows that still demand premium US tools.

Frontier intelligence is becoming more capable, more available, and less exclusive. That should expand adoption, but it also makes raw model access a less reliable source of durable advantage and shifts value toward the assets and capabilities that make intelligence useful, governable, and hard to replace.

AI is shifting from a tool that generates responses to a tool that performs useful work. Recent model releases have improved reasoning, reliability, memory, and the ability to work across different types of data. Anthropic’s AI model, Claude, helped accelerate the move from coding assistant to autonomous coding agent through Claude Code, then extended that logic into broader knowledge work. While this is a major disruptive theme and a central focus of this paper, it is only one part of a broader transformation. Specialist models in mathematics, biology, and other fields are proliferating, while researchers continue to experiment with non-transformer architectures such as world models designed for physical AI. Smaller models are improving as well, pushing more inference to run on local systems instead of keeping them entirely in centralized cloud environments.

These advances are beginning to make disruption more visible. In software, faster model improvement is compressing product cycles, narrowing functional differentiation, and pressuring application-level moats that once appeared durable. In labor-intensive workflows such as customer service, the effects are already showing up in shorter handle times, lower staffing needs, and greater pressure to automate routine work. Much of the broader disruption still lies ahead, but the direction is clearer.

Enterprise AI adoption is rising, but scaled deployment remains limited

The pace of enterprise adoption is becoming increasingly apparent, even if scaled deployment remains limited and difficult to track in real time. McKinsey survey data captures the key point for investors: adoption is rising, but much of it still reflects experimentation, piloting, and limited deployment rather than full integration across core workflows.

Coding remains the clearest use case, though productivity gains are still early and uneven. Deployment inside real organizations will be challenging, requiring integration, oversight, and governance as much as good model performance. Usage will evolve as service models change. For example, firms are still enjoying subsidized model pricing, internal controls are weak, and many firms are still learning to use these tools effectively while keeping token costs under control.

As AI advances rapidly, the need for stronger corporate governance is becoming more urgent, even if progress is moving at a more human pace. Some early adopters are pulling ahead in part because they addressed governance, access controls, and oversight sooner, making scaled deployment in sensitive workflows easier. Others may appear to be moving faster precisely because they are deferring those disciplines and accumulating a governance backlog that has not yet surfaced in operating results. Recent events also show how quickly regulatory and security concerns can affect commercialization. Anthropic’s temporary withdrawal of Fable 5 and Mythos 5 following a US government directive serves as a reminder that deployment risk may increasingly hinge on security, liability, and policy judgments. Public backlash is also growing, which could make regulation more political over time. As AI use broadens, questions of data provenance, auditability, security, and liability are likely to matter even more.

This shift from experimentation to scaled use is changing the investment question. Technical progress and broader adoption are making AI more commercially relevant, but they do not by themselves determine where durable returns will accrue. That depends increasingly on the economics of deployment and on which firms can turn AI capability into repeatable business action.

How AI economics are evolving

In our last edition of VantagePoint: The Rearview Mirror Problem, we argued that investors often mistake recent winners for future return drivers. That risk is especially acute in AI. The first-wave beneficiaries are well known, and the infrastructure buildout has become the market’s central focus. The harder question now is how the economics are evolving beneath that narrative, and which parts of the opportunity set can still deliver durable returns. We think investors should focus on three underwriting questions: Which bottlenecks are durable? Can rising revenues justify the capital required to sustain leadership? Where can value persist as AI becomes cheaper, more capable, and more widely available?

The economics are changing along with the technology. Constraints have shifted from training toward inference, usage, and deployment, and value capture is broadening with them. The relevant opportunity set now extends beyond frontier-training hardware to a broader mix of inference and data infrastructure, deployment software, workflow and permission layers, governance tools, and applications that shape how AI is embedded in business processes.

Earnings have begun to catch up with enthusiasm in parts of the AI ecosystem. Recent gains in AI-linked equities no longer rest on expectations alone. Several leading firms have reported strong revenue growth tied to AI demand, especially in semiconductors, cloud, and selected infrastructure segments. Private model providers, such as Anthropic, appear to be seeing similar momentum, though their economics remain less transparent. Stronger fundamentals validate part of the move. They do not settle the harder question of whether revenue growth will prove durable enough to justify the operating and capital costs required to sustain it.

Which bottlenecks are durable?

This first question deals with whether bottlenecks are persistent or simply reflect temporary undersupply. Early in the cycle, scarcity centered on training compute and raw GPU capacity. That is no longer the full story. As AI deployment scales, the tighter constraints are becoming more physical. Data center infrastructure, memory, and advanced packaging remain important, but power is emerging as the clearest hurdle. Reliable electricity, cooling, transmission, and the ability to bring new capacity online increasingly matter as much as access to chips. ¹

Falling prices in one layer of the stack do not remove constraints in another. Listed token prices have fallen sharply, which should broaden adoption. But lower prices do not mean lower compute demand. As models have become more capable and AI systems take on more complex tasks, the compute required to complete useful work has continued to rise. More autonomous and always-on systems will reinforce that trend by increasing token consumption and placing greater strain on the physical stack.

Power deserves particular attention because it is increasingly one of the hardest constraints to relieve. Utilities, grid equipment, cooling, and related enabling infrastructure can be difficult to replicate quickly because they depend on permitting, transmission access, engineering capacity, and time to build. Those are more durable barriers than the temporary scarcity that can emerge in parts of the hardware stack early in a buildout.

Memory and advanced packaging also remain important choke points, supporting stronger pricing and earnings across parts of the semiconductor ecosystem. The key investment question, however, is not whether these areas are constrained today, but whether those rents are likely to persist. Some supply bottlenecks may prove temporary as capacity expands. Others may reflect capabilities that are harder to replicate quickly.

In some parts of the market, security, compliance, and regulatory approval may also function as bottlenecks. Where customers need trusted systems for sensitive workflows, firms that can meet higher standards for resilience, auditability, and control may accrue durable advantage.

Not every bottleneck supports durable economics. Some stem from short-lived pricing power. Others are tied to assets, regulation, siting, expertise, or customer relationships that are harder to reproduce. Open and lower-cost models reinforce that point. They may broaden adoption and accelerate experimentation, but they also challenge the idea that frontier capabilities alone guarantee durable pricing power. In areas where customers do not require frontier performance or tightly integrated proprietary systems, improving open models are compressing economics at both the model and software layer.

Can rising revenues justify the capital required to sustain leadership?

The second question asks whether rising revenues and earnings can justify the capital required to sustain AI leadership. That issue is now most visible in the capex cycle. For current spending to earn attractive returns, revenue growth must continue to catch up with investment, usage must remain high, enterprise monetization must deepen, and margins must hold up despite a much larger capital base.

Consensus expectations for the five major hyperscalers call for combined revenue to increase 54% while EBITDA is expected to increase about 111% from year-end 2025 through 2028. Over the same period, depreciation is expected to rise much faster. Depending on assumed asset lives, it could increase by roughly 175% to more than 340%. That drag is large enough to matter. At the low end of those estimates, 2028 depreciation would come close to the group’s 2025 net income of $405 billion.

AI is making important parts of technology more capital intensive. Some parts of the market may still be valued as if AI were reinforcing capital-light software economics, when in fact it is making important parts of the stack more asset-heavy and operationally demanding. Investors should place more weight on depreciation, reinvestment needs, financing conditions, and the durability of pricing power on these more asset-heavy companies.

Memory is not a direct analogue for hyperscalers, and the current AI cycle has different drivers. Still, its history is a useful reminder that periods of tight supply, strong pricing, and high margins can look more durable than they prove to be once capacity expands. Shortages have repeatedly lifted margins and encouraged new investment, only to erode those same margins as supply caught up. AI may not follow that path exactly, but the lesson is familiar. Strong demand does not by itself protect returns when supply can respond and pricing power is not well defended.

Indeed, Micron and SK Hynix—two of the memory companies most directly exposed to advanced AI demand—have increased capex by a combined 70% in each of the last two years, and consensus expects another roughly 55% increase in 2026. Investors should be careful not to assume that today’s strong pricing and profitability will persist unchanged as capital spending rises and supply responds.

The quality of demand matters as well. Investors should distinguish between durable end demand and demand supported by ecosystem-linked commercial arrangements, strategic subsidy, or circular deal structures that make near-term economics look stronger than they are. As the system matures, leverage and structured financing also deserve more attention. Risk rises when capital assumptions become aggressive ahead of proven cash flows. After rising by roughly $900 billion since the start of this year, gross supply of investment-grade credit is expected to increase by about 17% over the full year to a record $2.1 trillion, with much of the increase coming from hyperscalers and related infrastructure. Structured credit markets are expected to see data center securitizations rise by nearly 50% in 2026 to $30 billion.

Stronger fundamentals have made the buildout more credible, but not necessarily more durable. High depreciation expense creates a demanding hurdle for hyperscalers that are increasingly competing with one another for business. Capital-intensive businesses facing rising competition may struggle even if the addressable market continues to grow. Not all participants will fare well. Semiconductors and advanced memory should benefit from tight supply and, in some cases, multi-year contracts, but supply is likely to catch up over time as capacity expands and technology becomes more efficient. Investors should be cautious about extrapolating today’s pricing and profitability too far into the future.

Where can value persist as AI becomes cheaper, more capable, and more widely available?

The final question considers where durable value can persist. Access to models alone will not remain enough. As intelligence diffuses, we expect more defensible positions to belong to firms that control how it is used: who owns the workflow, governs permissions, controls distribution, and connects output to execution. Hyperscalers and other large platforms are trying to capture value across multiple layers of the stack through vertical integration, from compute and cloud infrastructure to model access, routing, deployment, and enterprise tooling.

The most important shift is in what software and adjacent systems actually do. As agentic systems begin to perform economically meaningful work rather than simply assist users, part of the addressable market shifts from software budgets to labor budgets, which are much larger. That may expand revenue pools, deepen integration, and create stronger business models. It may also intensify disruption across software, services, and selected consumer sectors.

Software and adjacent control layers may still be where much of the value ultimately accrues, but they also pose the hardest underwriting questions. AI may expand revenue pools even as it weakens traditional moats. Customers may expect broader functionality without proportional price increases, while model, compute, orchestration, and support costs remain material. We expect the stronger positions belong to firms that are deeply embedded in a workflow, possess privileged task-specific context, and can convert AI output into completed work rather than simply sell access to a feature.

That distinction matters because control over the workflow is different from access to the model. A firm that helps generate an answer may be easy to displace. In an agent-driven environment, durable advantage should rest increasingly on control over permissions, approvals, and execution rather than on data alone. Systems that determine what autonomous software can access, trigger, and complete could have a competitive upper hand over systems of record alone.

This logic extends beyond enterprise software. In consumer markets, such as commerce, education, and travel, AI is likely to reshape discovery, service, and execution. New products and business models should emerge. But these same layers may also face the greatest pressure from improving models and larger platforms, especially where functionality is easy to replicate, customer relationships are weak, or distribution is controlled by someone else.

The same logic also shapes investment underwriting. Strong adoption and fast top-line growth may not be enough if a company with limited bargaining powers depends heavily on a single model provider, hyperscaler, or distribution platform. Downstream growth may prove real without translating into durable economics. Strategic acquisition may become a common end state for promising firms, supporting investment outcomes alongside a select group of independent long-duration compounders. For private equity and venture investors, underwriting should place more weight on customer ownership, monetization after model and compute costs, governance quality, likely end states, and the durability of economics if acquisition interest fades.

Investment implications

AI should be treated as a system-wide set of exposures rather than a narrow thematic trade. We see the stronger opportunities ahead in harder-to-relieve bottlenecks—especially power and related infrastructure—alongside the software and application layers that govern deployment in real workflows and emerging AI-native businesses that can reshape industry economics. The same shift should benefit adopters that use AI to improve their own economics while increasing disruption risk for incumbents that fail to adapt or are displaced by new business models.

This shift argues for more caution toward parts of the AI ecosystem where expectations, capital spending, and competition have all risen sharply at once. Large hyperscalers remain central to the buildout and may continue to benefit from scale, distribution, and enterprise integration. But they are also engaged in an increasingly costly race to secure compute, power, and physical infrastructure, and the associated depreciation burden is becoming harder to ignore. We therefore lean away from the most crowded first-wave winners, particularly where valuations still leave limited room for disappointment. The same caution applies to parts of semiconductors and memory, where recent earnings strength has been real, but history suggests investors should be careful not to mistake tight supply and current pricing power for durable advantage.

By contrast, we are more constructive on select infrastructure and real assets tied to harder-to-relieve constraints, particularly electricity infrastructure, grid access, and related enabling assets. These areas appear better positioned to benefit as AI deployment scales and physical bottlenecks become more binding.

The widening opportunity set also creates room for emerging disruptors, many of which were inconceivable before recent AI advances. As AI becomes cheaper, more capable, and more widely available, value will migrate toward firms that control workflows, permissions, customer relationships, and operational integration rather than those relying on thin wrappers or temporary model arbitrage. In software, the more durable positions are likely to belong to companies that can embed AI into economically meaningful tasks, govern it effectively, and monetize completed work rather than simple access. Private equity may benefit where businesses need capital, operational support, and technology investment to adapt successfully to AI-driven changes in cost structure and competition.

Venture capital remains the key channel for accessing emerging AI-native companies and disruptive business models, and investors seeking that upside likely need some participation in private markets. But this technology cycle is still early and, as in past cycles, a few winners are likely to emerge alongside many losers as innovation advances faster than commercial adoption. In that environment, exposure to AI is not the same as access to strong investment returns. Because private commitments are long-lived, pacing matters as much as manager selection. Investors should continue to allocate selectively, with discipline on timing and valuation rather than rushing to add exposure simply because the theme is compelling. As a new wave of highly anticipated technology IPOs comes to market this year, investors should be thoughtful about redeploying capital to venture capital, balancing those opportunities against other market segments with more attractive valuations and differentiated return potential.

The same framework should also be applied defensively. AI is not only a source of new opportunity. It is also a source of disruption risk in existing holdings. Businesses with weak differentiation, labor-intensive models, or information-heavy processes may be more vulnerable than they appear, even if they sit outside any obvious AI category. Equity long/short hedge funds may be well positioned to benefit from rising dispersion as AI creates clearer winners and losers across software, services, and other information-intensive industries. As disrupted companies—especially ones that took on private debt during the period of zero interest rates and high valuation from 2021 to early 2022—struggle to refinance over the next few years, stressed and distressed opportunities may emerge.

AI remains an important area of exposure. From here, we expect the best opportunities to come from identifying durable bottlenecks, defensible control points, and the businesses most likely to benefit from disruption rather than suffer from it. Active management across public and private markets will be central to success.

Graham Landrith and Justin Hopfer also contributed to this publication.

Index Disclosure

MSCI All Country World Index (ACWI)
The MSCI ACWI captures large- and mid-cap representation across 23 developed markets (DM) and 24 emerging markets (EM) countries. With 2,558 constituents, the index covers approximately 85% of the global investable equity opportunity set. DM countries include Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Hong Kong, Ireland, Israel, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Singapore, Spain, Sweden, Switzerland, the United Kingdom, and the United States. EM countries include Brazil, Chile, China, Colombia, Czech Republic, Egypt, Greece, Hungary, India, Indonesia, Korea, Kuwait, Malaysia, Mexico, Peru, the Philippines, Poland, Qatar, Saudi Arabia, South Africa, Taiwan, Thailand, Turkey, and the United Arab Emirates.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post VantagePoint: Artificial Intelligence Investing After the First Wave appeared first on Cambridge Associates.

Today’s market leaders remain strong businesses, but they are also priced for a continuation of unusually favorable conditions. Meanwhile, valuation gaps across regions and styles are wide; inflation risks are less settled; and the geopolitical, policy, and fiscal backdrops are less benign than investors had come to expect. In this edition of VantagePoint, we explain why investors should be wary of relying on rearview mirror assumptions, where current concentrations create the greatest vulnerability, and why more attractive opportunities now lie beyond the market’s recent winners. The implication is not to react abruptly, but to act strategically.

Investors are not prepared

At the end of every bull market, investors tend to be overallocated to whatever worked best. Today, that means not only US equities and US dollar assets, but also a US market that has become unusually concentrated. The United States now represents roughly 64% of the MSCI All Country World Index, up from about 42% in 2010, and the top 10 US companies—most of them technology related—account for 24%. The information technology sector is near its 36% share of the US market reached earlier this year, exceeding the concentration seen in the late 1990s, and that doesn’t even include Amazon, Alphabet, Meta, and Tesla, which are classified as consumer discretionary and communication services stocks. Foreign capital has been attracted to US assets, supporting the US dollar. As was the case in the 1990s, once the enthusiasm for US equities fades, the US dollar will likely fall as well.

US household equity ownership has also risen to unusually high levels relative to net worth, with equities exceeding real estate by the widest margin in the post-World War II era. Previous episodes in which a narrow group of companies came to dominate market returns—including the Nifty Fifty era in the late 1960s/early 1970s and the dot-com era of the 1990s—proved poor moments to abandon diversification. The point is not that these exposures are about to collapse. It is that many portfolios now rest on assumptions that have become less reliable.

Most investors have correspondingly little direct exposure to tangible assets with real-world scarcity value. That mix made sense in a world shaped by falling inflation, expanding globalization, and stable supply chains. It makes less sense in a world where supply constraints, geopolitical friction, and higher capital intensity are becoming more persistent features of the investment landscape.

The supply side of the economy has changed in ways that look more structural than cyclical. In just five years, investors have had to absorb a series of major shocks: the COVID-19 pandemic, Russia’s invasion of Ukraine, US tariffs, and now the Iran War. Each has reinforced the same lesson. Inflation is not simply a demand-management problem, and geopolitical risk is not a remote tail event. It is a recurring influence on growth, prices, capital flows, and investment priorities. Geopolitical rivalry, defense modernization, artificial intelligence (AI) adoption, energy security, and climate change mitigation and adaptation are increasingly moving together, with important consequences for capital spending, inflation, and asset returns.

This is the rearview mirror problem. Investors are still using the conditions that shaped the past 15 years as a guide for what comes next, even as that backdrop becomes less dependable. We made a similar point in early 2000, when we warned that apparent diversification had been undermined by a shared dependence on US technology exposure. The parallel is not exact, but the lesson is familiar: portfolios built around what worked in one regime can become more fragile than they appear when conditions change.

AI is eating the world, at a cost

AI is the most powerful expression yet of a much longer cycle of US tech leadership. What began as a market preference for scale, duration, and capital-light growth, evident in the earlier dominance of the FANGs (the popular acronym for Facebook, Amazon, Netflix, and Google first coined in 2013) has evolved into a dependence on a small group of companies tied, directly or indirectly, to the AI buildout. Their influence now extends well beyond equity benchmarks. AI-related spending is helping to drive corporate capex, support economic growth, shape investment-grade debt issuance, and increase demand for electricity generation, transmission, and grid resilience. As in past booms, a single theme has grown large enough to shape multiple parts of the investment landscape at once.

That broader reach is one reason historical parallels are useful. Major technological revolutions have repeatedly followed a familiar path. Railroads, electrification, and the internet all produced real economic transformation. They also produced overbuilding, excessive optimism, and eventually some form of bubble dynamic. That is not a contradiction. It is often how transformative technologies are financed. Capital rushes toward the most visible opportunities, investors assume early leaders will capture most of the value, and markets discount a future that proves harder, more competitive, and more capital-intensive than expected. There is little reason to expect AI to be different.

The late 1990s offer a particularly useful comparison. The problem then was not simply speculation. It was that market leadership, capital spending, financing activity, and investor expectations became tightly bound to a single secular narrative. Many dot-com companies were highly speculative and unprofitable, with Pets.com serving as a useful poster child for the era. But companies like Cisco, Microsoft, and Oracle were real companies with strong businesses and central roles in the digital economy. The lesson is that even genuine technological leaders can become poor investments when valuations imply too smooth a path from innovation to durable returns. The same risk exists today.

Recent market weakness has taken some of the air out of the most stretched valuations, although much of the improvement has been retraced in April’s sharp market rally. Forward earnings multiples for the Magnificent 7 are meaningfully below their peaks. And on sales-based measures, the derating has been less pronounced, suggesting that investors are still paying demanding prices for a narrow set of companies expected to deliver an unusually large share of future growth. Some froth has come out of the market, but it remains heavily dependent on continued execution from the same leadership cohort.

The financing side of the story deserves at least as much attention as valuation. Hyperscaler capex has risen sharply even as free cash flow has started to decline from 2024 peaks, narrowing the gap between internally generated cash and the spending required to maintain leadership. Indeed, these companies are expected to spend $700 billion on capex in 2026—a 70% increase over 2025 spending, which will eat into free cash flows even as operating cash flow is still rising.

As that gap closes, debt financing—including off-balance-sheet structures—is becoming more important. The breadth of the theme is visible in private markets as well. AI and machine-learning deals accounted for roughly half the value of global venture capital investment in 2025 and are expected to account for an even larger share in 2026, up from about one-fifth in 2020 and almost nothing in 2010. That is another sign that AI is no longer simply one promising area of innovation. It is increasingly the organizing principle for capital formation across the growth ecosystem. That may create important opportunities, but it also reinforces the need for discipline. When one theme absorbs such a large share of capital, the line between durable advantage and speculative excess becomes harder to draw.

Just as important, the market may still be valuing some of these companies as if they retained the economics of capital-light software businesses. History suggests caution. Firms that grow assets aggressively have tended to underperform more capital-efficient peers, a pattern consistent enough to be embedded in academic factor models. If AI leadership increasingly requires sustained investment in data centers, chips, power, and networks, some of today’s leaders may deserve a different valuation framework than the one investors have become accustomed to applying. Consider that if current analyst estimates are correct, the hyperscalers will have accumulated $2 trillion in AI-related assets by 2030. Assuming an average life of five years, or 20% depreciation, would result in $400 billion in annual depreciation charges, roughly equaling their combined profits of $405 billion in 2025. The AI cycle is moving from enthusiasm to financing, and markets may still be underestimating the eventual cost of leadership.

The investment risk, then, is not simply that AI enthusiasm has gone too far. It is that many portfolios are more dependent on a single theme than they appear. When one secular story drives equity concentration, capital spending, credit issuance, infrastructure demand, and venture enthusiasm all at once, the case for diversification becomes stronger, not weaker. Investors do not need to reject AI to recognize that the better long-term opportunity may lie in markets where expectations are lower, valuations are less demanding, and portfolios are less dependent on a single story.

Credit as an early warning system

Credit deserves attention because it often reveals fragility before equity markets do. The signal today is mixed. Traditional default rates in public credit have eased, and reported defaults in private credit appear to have eased as well. But broader measures that include distressed exchanges, liability-management exercises designed to avoid default, and payment-in-kind restructurings paint a less comfortable picture. Headline default rates still look manageable, but they may understate where strain is building. That is especially true in private markets, where quarterly marks, amend-and-extend activity, and abundant capital can delay recognition of weakening credit conditions.

AI adds another layer to this picture. As the financing cycle has progressed, more of the capital required to support AI-related investment has moved beyond equity enthusiasm and into credit markets. Investment-grade credit is expected to see gross supply rise about 25% this year to a record $2.25 trillion, with a 10x increase to an estimated $400 billion coming from hyperscalers and related infrastructure. Structured credit markets are expected to see data center securitizations rise by nearly 50% to $30 billion. That includes financing for data centers, infrastructure, and businesses whose economics remain unsettled. It also matters for software-heavy loan books, where some of the most aggressively structured deals were made in businesses that now face greater competitive pressure or pricing uncertainty as AI diffuses. Parts of the market are now being tested against assumptions formed in a more benign period.

Credit market structure has also changed materially over the past decade, with important implications for who provides financing, where leverage sits, and how stress could spread through the system. Private credit is now a much larger and more influential part of the financing ecosystem than it was a decade ago. Much of that growth reflects tighter bank capital regulation after the Global Financial Crisis (GFC), which made some forms of lending less attractive for banks and created room for private lenders to expand. Investor capital followed, drawn by higher yields and the promise of illiquidity premia. This shift brought real benefits, including broader access to capital and more flexible financing for some borrowers. But it also intensified competition, particularly in direct lending, where spreads are tight and lender protections have weakened in more crowded parts of the market.

The same shift was reinforced by the long period of near-zero interest rates that followed the GFC. Zero Interest Rate Policy (ZIRP) did not just lift asset prices. It also encouraged financing structures and underwriting assumptions that were easier to sustain when capital was cheap, and refinancing was routine. Highly levered companies, aggressive growth strategies, and buyouts struck at elevated multiples all looked more manageable in that environment. Many businesses will prove less resilient in a higher-rate world, especially where earnings growth is slowing, equity cushions are thinner, and valuations remain anchored to a more forgiving era. The clearest pressure point is the cohort of loans originated in 2021, when financing terms were exceptionally easy, equity valuations were overstated, and leverage was often pushed to levels that are harder to refinance today. That matters for private equity as well as private credit and secondary funds that are picking up these companies as they come to market. Some of the most vulnerable credits are direct loans tied to sponsor-backed transactions completed when financing was abundant.

The main risk is not necessarily systemic in the way investors associate with 2008. But the warning is still meaningful. Pockets of strain in credit reinforce the broader message of this note: portfolios built around a narrow set of favorable assumptions may be more fragile than they appear. A slowdown in private credit lending would still amount to a form of credit contraction, potentially weighing on growth and certainly increasing the cost of the AI buildout. Risk to the banking system is limited for now, although bank exposure to the private-credit ecosystem has grown and available data almost certainly understate the full extent of those linkages. Insurance is another area to watch, particularly, if private-letter credit ratings overstate underlying credit quality. In such circumstances, highly levered, thinly capitalized insurance companies that have accumulated too many direct loans may come under pressure. Recent concern has also centered on semiliquid vehicles that offer more liquidity than the underlying private credit assets they hold. These funds have seen increased retail outflows, raising the possibility that stress may emerge through gating, valuation uncertainty, or a more selective and uneven availability of credit rather than through a single market-wide break.

For investors, the takeaway is not to avoid credit, but to recognize that portfolios built with disciplined manager selection, underwriting, and fund structuring during more exuberant periods should be better positioned to navigate market shifts. Diversification across sub-strategies should also provide ballast to portfolios.

Where to look for diversification

Investors should respond by trimming crowded exposures and rebuilding diversification. The most compelling opportunities now lie in areas where valuations are lower, expectations are less demanding, and return drivers are less tied to the same crowded narrative. That points first to non-US equities, value, small-cap equities, active strategies including hedge funds, and real assets tied to a more capital-intensive and electrified economy.

The strongest public market opportunity is outside the United States. Non-US equities offer lower valuations, less concentration, and greater exposure to sectors and styles left behind during the long period of US large-cap dominance. A weaker dollar, which we expect, would provide an additional tailwind for non-US equity exposure. The combination of valuation support and currency tailwind has historically been a powerful setup for extended periods of outperformance. Global ex US equities are also tilted to traditionally value-oriented sectors, which adds to their appeal in the current environment. The broader capex cycle now underway, while still tied to AI, also incorporates more geographically dispersed themes of energy security, grid resilience, and defense, providing additional support to non-US markets with deeper exposure to industrials, utilities, and related cyclicals.

Small-cap equities also look more attractive than they have for some time. Relative to large caps, valuations are modest, and earnings expectations are beginning to improve as growth broadens beyond technology-related winners. Small-cap equities, especially in the United States, have been held back by higher financing costs, weaker balance sheets, and the market’s overwhelming preference for scale. But that is also why they offer greater upside if capital becomes more selective and market leadership broadens.

Near-term risks remain. Global ex US equities and global small caps are more exposed than US large caps to economic disruption tied to conflict in the Middle East, and the dollar could strengthen further in the short run. But the situation is too fluid to time tactically. Investors are better served by rebuilding diversification and using periods of renewed dollar strength and US equity outperformance to add to non-US positions and reduce US dollar exposure.

A less concentrated market would also improve the opportunity set for active management. The dominance of large-cap equity performance has created a powerful headwind for active managers. Broadening market leadership and ebbing concentration would change that. Within hedge funds, the case is strongest for strategies that benefit from greater dispersion rather than broad market direction. Equity long-short managers should have a better opportunity set in a world where valuation matters more and returns become less concentrated. This may be particularly fruitful in less efficient markets outside the United States. At the same time, less directional hedge funds, such as arbitrage, global macro, and trend-following strategies can help diversify portfolios when stock-bond relationships become less reliable and macro shocks reverberate across markets in less predictable ways. Other diversifying strategies, such as insurance-linked securities and asset-backed credit can also provide diversification. These are not perfect hedges, but they are better suited to a more fractured environment than portfolios that rely on cash and sovereign bond duration as their only ballast.

Real assets should also play a larger role, especially the resources and infrastructure needed to support a more capital-intensive and electrified economy. Electricity infrastructure and grid modernization are among the most compelling structural opportunities in real assets. The International Energy Agency estimates that $600 billion per year in grid investment is needed by 2030, roughly double current levels. The AI buildout, reshoring, and the energy transition are all increasing demand for generation, transmission, storage, and grid efficiency. The Iran War reinforces the strategic value of energy security, resource independence, and AI-related capabilities, all of which are increasingly being treated as matters of national security. These are long-duration needs with real economic importance and, in some cases, attractive supply/demand characteristics. For portfolios that have become too dependent on financial assets and intangible growth, this is one of the clearest ways to rebuild exposure to scarcity value and real-economy investment.

Commodities and natural resources offer some inflation sensitivity and exposure to supply constraints, though timing has been difficult, particularly given weakness in the Chinese economy and property sector. Years of underinvestment in mining and extraction have created meaningful supply deficits in copper, lithium, nickel, and uranium. The structural case is strong, but current pricing already reflects some of that scarcity even as Chinese demand remains soft. Should the tail risk scenario of the Iran War push the global economy into recession, valuations in parts of the complex would likely become more attractive, creating better entry points and reducing the timing risk that has frustrated investors in this space. Investors should approach this opportunistically, building positions when prices allow rather than chasing what has already moved.

Private markets also require more discrimination. Venture capital is likely to produce some of the future winners in AI, and investors who want exposure to that upside will need some exposure to the private ecosystem. But this is still an early-stage technology cycle, and the eventual winners are far from settled. Investors should be careful not to confuse access to the theme with access to the returns. Meanwhile, many existing technology investments, especially enterprise software deals struck when rates were near zero and valuations were generous, are being tested in a more demanding environment of higher rates and AI disruption. Because private commitments are long-lived, pacing matters as much as manager selection. Investors should continue to allocate to private investments selectively, with patience about timing and valuation rather than rushing to add exposure simply because a theme is compelling. If a new wave of eagerly anticipated technology initial public offerings (IPOs) comes to market this year (e.g., SpaceX, Anthropic), investors should be measured in redeploying capital to private investments, balancing those opportunities against other parts of the market offering more attractive valuations and differentiated return streams. Especially in AI- and software-focused private equity, investors should favor managers with valuation discipline, rigorous underwriting, and the ability to distinguish durable advantage from enthusiasm financed on easy terms.

In short, portfolios built with too much concentration in prior winners are increasingly fragile. We cannot know precisely when leadership will change, but we do know that portfolios are more resilient when they are not built on the assumption that the recent past will continue indefinitely. We believe the strategic response is clear: trim crowded exposures and rebuild diversification through non-US equities, value, small-cap equities, real assets, and active strategies that can benefit from greater dispersion. Investors do not need to abandon the market’s recent winners. They do need to stop treating them as the only place to look for durable long-term returns.

Drew Boyer and Justin Hopfer also contributed to this publication.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post VantagePoint: The Rearview Mirror Problem appeared first on Cambridge Associates.

That environment is now changing. Weakening underwriting standards, retail vehicle growth, and technological disruption are creating a more challenging landscape. Given these changes, we expect performance dispersion to emerge more clearly between managers that remained disciplined throughout these heady times and those that did not. This widening range of outcomes reinforces the importance of manager selection and taking a diversified approach across strategies, geographies, and borrower segments.

How we got here

Since the Global Financial Crisis, banks have steadily retreated from many traditional lending markets, creating an opportunity for private lenders. Over the past five years alone this trend accelerated, with US direct lending AUM roughly doubling to $1.3 trillion. Retail-oriented vehicles like interval funds and business development companies (BDCs) have played a key role in this recent expansion, with BDC assets alone now totaling more than $500 billion.

The rapid growth of assets, particularly via semi-liquid funds, has coincided with a slower merger & acquisition (M&A) environment, leaving more capital competing for fewer deals. The result has been intensified competition among direct lenders and a borrower-friendly market characterized by spread compression, weaker underwriting standards, and looser lender protections across the upper, core, and lower middle market.

Meanwhile, the benign credit environment that prevailed for much of the past decade—marked by historically low default rates—masked differences in manager quality. Because direct lending returns are highly asymmetric and driven largely by loss avoidance and recoveries, weak underwriting or limited workout experience often did not show up in returns. As a result, the gap in total value to paid-in (TVPI) between top- and bottom-quartile managers narrowed to just 0.1x, understating the true dispersion in underlying portfolio quality (Figure 1).

Making sense of the negative headlines

Since late 2025, the press has highlighted valid risks in private credit, but the coverage often exaggerates the extent and, thus, vulnerability.

Retail vehicles and asset-liability mismatch

Retail-targeted vehicles offering limited quarterly liquidity have created a mismatch between investor liquidity expectation and the illiquid, five- to seven-year nature of the underlying loans. Private BDCs manage this tension by capping quarterly redemptions at 5% of net asset value through discretionary tender offers, but recent increases in withdrawal requests have led some managers to activate these gates, weighing on sentiment (Figure 2). This reaction, however, appears partly driven by investor (and perhaps media) misunderstanding of the fund structure and its intended liquidity profile.

Gating can protect investors by preventing fire sales or the selective liquidation of high-quality assets that would disadvantage remaining investors. This risk is amplified by the use of leverage in BDCs, as deleveraging can generate further pressure on asset prices. Managers with healthier loan portfolios and sophisticated liability management are better positioned to manage through these pressures. In contrast, managers facing sustained redemption pressures, credit deterioration, and weaker liability management may be left holding more stressed or illiquid assets, increasing the risk of losses.

Institutional investors in longer-duration, lock-up vehicles are not directly exposed to these dynamics. Near-term spillover from retail outflows may affect deal flow and slow deployment, but the longer-term implications could be constructive. If redemption pressure persists, the supply/demand dynamic may become more balanced, positioning managers with stable capital bases to deploy into opportunities with wider spreads and stronger lender protections.

Weakening underwriting standards

An oversupply of capital relative to deal opportunities has pressured direct lending, compressing yields and weakening underwriting standards and lender protections. This has been most evident in the US upper middle market, where large publicly traded asset managers have concentrated their retail-oriented vehicles. This segment began to attract a disproportionate amount of direct lending AUM in 2022, when Federal Reserve rate hikes and recession fears dislocated public loan markets and created attractive pricing for direct lenders. However, once the broadly syndicated loan market reopened, the direct lenders in the upper middle market began competing directly with the public markets (Figure 3).

That competitive overlap has weakened deal terms. To compete with the broadly syndicated loan market, the upper middle market has long accepted the absence of financial maintenance covenants. Competitive pressures have also weakened lender protections in parts of the core middle market, including fewer maintenance covenants and looser documentation in some transactions. Even where covenants remain, aggressive EBITDA adjustments, unrealistic financial projections, and loose definitions often weaken them and create a false sense of protection (Figure 4). By contrast, lower middle market and specialty lenders—which typically serve more complex borrowers—have largely maintained covenant discipline.

The supply/demand imbalance in the market has put newer and less-established entrants at a disadvantage. Managers with established sponsor and borrower relationships have greater access to repeat financing opportunities, while less-established entrants are more likely to compete in broadly marketed transactions where documentation and pricing discipline were weaker. In a strong economy with low interest rates, these underwriting decisions may matter less, but if rates stay persistently high or the economy slows down, these managers are likely to face elevated portfolio stress.

Early signs of stratification are already visible. While average metrics for the direct lending universe such as interest coverage ratios and non-accrual rates appear stable, the tail of weaker portfolios is growing (Figure 5). For example, 11% of public BDCs receive more than 10% of total investment income in payment-in-kind (PIK), while the average BDC receives only 6% of income in PIK.

Software exposure and AI disruption

Technology-related deals, and software in particular, have dominated leveraged buyout (LBO) activity over the past decade. Direct lenders have financed a significant share of these transactions, often at elevated valuations and leverage levels. On average, direct lenders carry approximately 20% exposure to the software sector, much of it originated during an era of low interest rates and optimistic growth assumptions. Additionally, a notable subset of software loans consists of annual recurring revenue loans: financing extended to companies with recurring revenue but little or no positive free cash flow. As these loans approach maturity, borrowers face a more difficult exit environment, marked by compressed valuations and tighter financial conditions, which raises refinancing and repayment risk.

Artificial intelligence (AI) poses two primary risks to software companies: obsolescence risk, as customers build proprietary solutions in-house, and margin compression, as lower-cost competitors erode pricing power and force incumbents into defensive investment. Companies with proprietary data, embedded workflows, and regulatory moats are better insulated, and some incumbents may benefit by successfully integrating AI into their products. Nevertheless, the full effects of AI disruption will take time to emerge, and uncertainty itself warrants caution.

Most software deal flow in direct lending has been concentrated in the upper middle market, precisely where aggressive lending practices and weaker protections are most prevalent (Figure 6). As a result, private software loans carry, on average, more than a full additional turn of leverage relative to other major direct lending sectors and exhibit lower interest coverage ratios. These metrics are often measured against adjusted EBITDA figures, which further overstate borrower resilience and understate underlying liquidity risk. In this environment, direct lenders that have emphasized strong creditor protections and disciplined deal structures appear best positioned. These features provide managers with greater control and flexibility, enabling earlier intervention when signs of stress arise. In addition, many of these managers have maintained lower software exposure than the broader direct lending market, reflecting a disciplined avoidance of the most competitive deals. Some borrowers may be forced into expensive capital solutions, disadvantaging equity and remaining credit investors.

Expect wider dispersion

The era of uniformly strong, low-dispersion direct lending returns is ending. Weaker underwriting in parts of the market, retail vehicle stress, and AI-driven uncertainty in software are converging to create a more challenging environment. Going forward, we believe outcomes will depend more on manager quality, documentation discipline, and sector selectivity than on market exposure alone. Investors should expect a wider gap between top- and bottom-quartile performance, making manager selection more consequential than at any point in the asset class’s recent history.

As previously noted, we continue to favor strategies like lower middle market and specialty lending. These segments remain less crowded and, in our view, better positioned to hold up in a tougher environment. We believe manager outcomes will increasingly depend on their ability to:

• Target less competitive market segments. In the United States, crowding in the core and upper middle market has compressed spreads and weakened lender protections, we believe making the less-trafficked lower middle market more attractive on a risk-adjusted basis. Managers with strong sourcing capability and the ability to underwrite complexity are likely to be rewarded, including specialty lenders that focus on non-sponsored borrowers, companies in transition, and out-of-favor sectors.
• Demonstrate disciplined underwriting and experienced workout capabilities. In a more challenging environment, capital preservation will depend on proactive portfolio management, early identification of stress, and the ability to maximize recoveries when credits deteriorate. In our experience, these capabilities are often, though not exclusively, associated with larger, more experienced managers that have invested through prior credit cycles.

While the focus of this paper has been on the US market, we’d note that the market structure in Europe is generally more favorable for direct lending investors. European direct lending remains less crowded and more disciplined, and it has not seen the same growth in retail-oriented vehicles that has fueled the excess capital, spread compression, and underwriting deterioration in parts of the US market.

Conclusion

After a decade of strong and stable returns, the direct lending market is entering a more demanding phase. Compressed spreads, weaker underwriting in parts of the market, retail vehicle pressures, and AI-driven disruption in software are collectively widening the gap between managers that remain thoughtful and disciplined, and those that do not. Investors who prioritize underwriting discipline, creditor protections, capital preservation, and thoughtful portfolio construction—including diversification across managers, geographies, and borrower segments—will be best positioned for what lies ahead.

Abby Coleman – Abby Coleman is an Investment Director for the Credit Research Group at Cambridge Associates.

Frank Fama – Frank Fama is Head of Global Credit Research and a Managing Director at Cambridge Associates.

Walker Haymond – Walker Haymond is an Associate Investment Director for the Credit Research Group at Cambridge Associates.

Wade O’Brien – Wade O’Brien is a Managing Director for the Capital Markets Research team at Cambridge Associates.

Graham Landrith also contributed to this publication.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post A New Era of Dispersion in Direct Lending Favors Disciplined Managers appeared first on Cambridge Associates.

Tight credit spreads are attracting issuers but offer little protection to investors if AI optimism fades or geopolitical risks, such as recent events in Iran, trigger a “risk-off” environment. US high-yield bond spreads of 289 basis points (bps) sit in the bottom decile of observed values. Similarly, although US investment-grade corporate spreads have risen slightly since reaching post-GFC lows in January, they remain in the lowest quartile on record. Spreads in adjacent markets also reflect stretched valuations: for example, commercial mortgage-backed security (CMBS) spreads have dramatically tightened over the last two years despite rising delinquencies.

Accelerating AI-driven capex could boost supply and put pressure on spreads. According to Morgan Stanley, gross US investment-grade supply will rise approximately 25% to a record $2.25 trillion in 2026, driven in part by hyperscaler and related infrastructure issuance rising to $400 billion—roughly 10x what they raised in 2024. Structured credit markets will also see surging supply, with data center securitizations expected to rise nearly 50% to more than $30 billion. Issuance is already well underway; Alphabet and Oracle together issued nearly $60 billion of new debt in February alone.

The financing of this AI-driven buildout will span the capital structure. Senior unsecured bonds will be the primary vehicle for large, investment-grade issuers. However, asset-backed securities (ABS) and project finance debt are increasingly being used, especially for data centers. The ABS market is seeing innovation as data center cash flows are securitized to tap new pools of capital. As financing structures evolve, investors need to ensure they have a clear picture of leverage and risk. In some data center deals, for example, tenants effectively backstop the project, yet the transaction nonetheless sits off balance sheet, obscuring its risk profile.

The potential mismatch between the useful life of AI-related assets and the maturity of the debt used to finance them also warrants close attention. While demand for data center computing resources is currently robust—for example, waitlists for AI chips and capacity are common—future data center demand is uncertain. If AI models become more efficient or if a technological leap reduces the need for processing power, chips and the buildings that house them could become obsolete before their debt is repaid. This would threaten a variety of different debt instruments, including asset-backed deals that finance chip purchases, securitizations backed by data center revenue, and unsecured debt. The fiber optic buildout of the early 2000s offers a cautionary parallel: overbuilding led to years of excess capacity and financial distress for some issuers, though the long-term utility of fiber ultimately proved out.

Strong fundamentals help offset some of these risks. Many hyperscalers have high (AA or AAA) credit ratings, reflecting low leverage, strong free cash flow, and ample liquidity, making them well-positioned to absorb additional debt. However, not all players in the AI ecosystem generate such healthy operating cash flows, and even those that do will see capex demanding a growing share of this over the next several years. Utilities and REITs that own and operate data centers are taking on significant leverage to fund expansion but often start with different fundamentals. Utilities, for example, often carry BBB ratings, and some data center operators have “junk ratings,” a sticking point in recent financing deals.

Historically low credit spreads, in our view, do not sufficiently compensate investors for current risks, which include obsolescence and rising structural complexity. Given these challenges, credit exposure should remain within policy targets, and higher-quality structured credit, such as agency-backed mortgage securities, should be considered as a substitute for corporate bonds that offer only marginally higher yields but greater vulnerability to cyclical and technological risks. Investors in AI-related credits should focus on high-quality issuers with strong balance sheets and stable revenue streams. Complexity should be accepted only when compensated by a meaningful premium.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post Should Credit Investors Be Concerned About Rising AI-Related Debt Issuance? appeared first on Cambridge Associates.

by TJ Scavone

Yields on most major developed market (DM) sovereign bonds reached a multi-year high in 2023 and have since held just below those highs, trading in a relatively narrow range. We expect this pattern to persist into 2026, supported by a resilient yet uncertain economic and policy backdrop, fair valuations in most markets, and ongoing yield curve pressures. Investors should keep allocations to high-quality sovereigns closely aligned with policy guidelines.

Looking ahead to 2026, the environment for most high-quality sovereigns remains broadly supportive. Economic growth is healthy but slowing—DM real GDP is projected to rise 1.7% in 2025, down from 1.9% in 2024, with most of the deceleration in the United States. While US consumer spending remains supportive, the labor market has softened, and the full impact of tariffs remains uncertain. These dynamics are likely to keep the Fed and other major central banks biased toward modestly easing in 2026, despite persistent inflation concerns. Overall, softer labor markets, tariff headwinds, and resilient but softer growth—supported by healthy consumer spending, AI capex, and easier policy—should limit both recession and inflation risks, resulting in modestly lower policy rates in many markets and rangebound sovereign bond yields in 2026.

Given this backdrop, we recommend maintaining exposure to high-quality sovereign bonds, with duration risk kept in line with benchmarks. The case for a short-duration stance has weakened as short-term rates have declined and yield curves have steepened, raising the opportunity cost of holding cash. Likewise, the case for adopting a long-duration stance is not compelling. Long duration typically outperforms when growth slows and central banks ease, but we anticipate only limited monetary easing. The European Central Bank and Bank of England have already delivered most of their anticipated cuts and markets are pricing in around 75 basis points (bps) of Fed cuts in 2026—a scenario that looks optimistic, considering current risks. Additionally, sovereign bond yields in key markets, like the United States and euro area, are currently in the bottom half of what we consider their fair value ranges, leaving little room for further declines absent a recession.

There are risk factors that warrant close attention. We have recently seen longer-duration sovereigns underperform as a range of influences—including fiscal concerns, elevated macro volatility, and cyclical factors—have put upward pressure on yields further out the curve. Fiscal pressures in particular have repeatedly made headlines in recent years, with many DM countries facing challenging fiscal outlooks and heightened volatility around budget stand-offs. While fiscal pressures warrant monitoring, market pricing does not signal imminent fiscal crisis, nor are they the sole driver. Elevated macro volatility, structural headwinds, and cyclical factors like monetary policy have also contributed. Many of these influences should reverse in a growth shock, allowing bonds to rally and provide portfolio ballast, as seen at points this cycle. However, with these crosscurrents, investors should demand more attractive yields before adding exposure. For context, yields would need to rise another 130 bps–180 bps to reach the upper end of their implied fair value range in the United States and Germany. Some regions offer more value, but domestic and currency risks need to be considered. In most cases, we recommend waiting for more attractive US Treasury valuations—given global spillover effects—before extending duration risk.

Overall, we anticipate that bonds will outperform cash in most major markets—supported by steeper yield curves—and should maintain their defense role in a downturn. However, since current bond yields are not especially attractive relative to our fair value estimates, we recommend maintaining allocations at policy levels, and keeping duration risk closely aligned to benchmarks.

Investors should underweight public corporate credit in 2026

by TJ Scavone

At present, the public credit universe offers few compelling opportunities. While returns have been solid and fundamentals remain sound, public credit is increasingly a one-sided trade. Spreads for both investment-grade and high-yield corporates are near historic lows, and the economic backdrop is turning less supportive. We see potential for spreads to widen in 2026 and beyond, and as a result, we favor higher-quality spread products that offer better relative value and more diversified return streams.

US investment-grade corporate bonds returned 6% annualized over the trailing three years as of November 30, and US high-yield bonds returned 10%. These strong returns were driven by high starting yields and a significant narrowing in credit spreads—down 52 bps for investment-grade and 179 bps for high-yield. The tightening in spreads, a pattern that was evidenced across most regions and instruments, was justified by robust economic and earnings growth, resilient corporate fundamentals, and subdued issuance, but yields are now less compelling, and spreads are historically tight across public credit.

While spreads could drift lower in the near term, upside for public credit is limited and downside risks have increased. The environment is more fragile, with slowing growth and emerging stress in the labor market and among low-income consumers and select corporate borrowers, highlighted by recent high-profile defaults. Riskier assets look increasingly vulnerable after the sharp run-up in equity valuations, as discussed earlier in this outlook, and the potential for slower growth and elevated costs could pressure corporate earnings and margins. Although material spread widening is not our base case, the credit cycle is maturing and risks favor wider spreads, supporting an underweight stance in public corporate credit within core fixed income.

Despite expensive public credit markets, select spread products offer compelling relative value. We favor US agency mortgage-backed securities (MBS)—particularly higher-yielding current coupons—and US municipal bonds (munis). We believe current coupon MBS are higher quality and well positioned to outperform if spreads widen, providing defense without sacrificing yield. Notably, current coupons (4.9%) now yield more than corporates (4.8%). Historically, at these levels, current coupons have outperformed corporates 62% of the time over the next two years, with returns ranging from -3% to 11% per year. Their spreads, unlike corporates, remain above historical lows with room to tighten as rate volatility subsides. Although rate volatility has declined since its recent peak, it remains somewhat elevated. With quantitative tightening ending and further modest rate cuts likely once tariff-related inflation pressures ease, there is scope for both volatility and MBS spreads to compress further, supporting returns.

Munis also offer attractive relative yields for taxable investors. For high-tax-bracket US families, munis have consistently delivered stronger after-tax returns than Treasury bonds and corporates. After adjusting for taxes, the yield advantage for munis is unusually wide—currently about 185 bps versus Treasury bonds and 93 bps versus corporates, among the widest taxable-equivalent spreads since the Global Financial Crisis, excluding isolated stress periods. Many taxable investors reduced muni holdings over the past decade, favoring Treasury bonds or, in some cases, even reaching for yield in credit, as low yields limited their tax advantage and valuations were less compelling. That is no longer the case, and the current environment favors shifting back toward munis at the margin.

Against this backdrop, it is important to recognize that public credit markets overall offer limited upside and heightened downside risk as spreads remain tight and the economic outlook softens. In this environment, we recommend a defensive posture within core fixed income, emphasizing higher-quality, more resilient sectors, with attractive relative value. US current coupon agency MBS and municipal bonds stand out for their relative yield advantage and diversification benefits. For those investors for whom these investments are appropriate, focusing on them may help position portfolios for more balanced risk-adjusted returns in 2026.

Investors should lean into private asset-based finance strategies in 2026

by Wade O’Brien

In 2026, credit investors face challenges such as expensive valuations, moderating growth and falling yields. Recent bankruptcies like First Brands and Tricolor also highlight the risk of weaker underwriting in at least some segments. We believe the solution is focusing on less correlated private credit strategies such as asset-based finance (ABF), insurance-linked securities, and litigation funding. Some of these strategies can be accessed via semi-liquid vehicles, freeing up illiquidity budgets for other parts of the portfolio.

Less correlated private credit strategies are attractive relative to expensive public credit assets. Strong demand has pushed spreads on assets like US high-yield and investment-grade bonds near the bottom decile of historical data, as we discuss elsewhere in this outlook. While demand across products is likely to be underpinned by yields near historical medians, returns are vulnerable if the pace of expected Fed cuts disappoints.

ABF funds offer investors the ability to diversify portfolios away from cyclical and expensive corporate lending. These funds lend against a variety of assets including consumer loans, real estate, and equipment leases. Underlying loans are less economically sensitive and have shorter maturities, allowing lenders to reprice them more quickly as conditions change. Accelerated cash return can also help investors concerned about slower distributions in other parts of their private portfolios. Recent bankruptcies have drawn attention to the ABF market, but were idiosyncratic, given the fraud and business practices involved. Still, they highlight the importance of careful manager selection, as both cases involved red flags that were ignored by markets. Fundraising by dedicated ABF funds has picked up but remains a fraction of the volumes seen in other private credit strategies.

While direct lending funds are currently less attractive in our view than less correlated private credit strategies, they remain attractive relative to comparable public credits. Fed rate cuts and lower spreads will impact returns, but fundamentals have been stable and defaults limited. The biggest near-term challenge for direct lending funds is competition from both the syndicated loan market and retail-targeted vehicles. Semi-liquid retail funds, including private business development corporations (BDCs) and interval funds, had accrued around $350 billion in assets by year-end 2024, a 60% increase in just two years. Reduced buyout volumes have cut supply and added to pressure on spreads, but resurgent M&A activity as rates decline and tariff uncertainty clears may help. Lower middle market lending funds, which offer higher spreads and better protections for lenders, are preferred to upper middle market.

Investors can access direct lending and ABF via open-ended vehicles as well as traditional closed-end funds. Private BDCs and interval funds may charge higher fees but offer investors the ability to more frequently adjust exposures. Investors that can access lower fee institutional evergreen funds may find them an attractive substitute for liquid credit assets featuring low spreads and yields.

Other private credit strategies—such as royalties, litigation finance, and insurance-linked securities—also have appeal. They tend to have resilient income streams insulated from the economic cycle and less sensitive to corporate fundamentals. Returns for these strategies have compared favorably with other types of private credit in recent years. These markets require highly specialized expertise, making their return streams less vulnerable to rising competition or surging demand from retail-targeted offerings.

In summary, with public credit markets offering limited value and increased competition, investors should look to private credit—especially ABF and specialized strategies—for better diversification, resilience, and risk-adjusted returns in 2026.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post 2026 Outlook: Fixed Income Views appeared first on Cambridge Associates.

Recent headlines have drawn attention to the bankruptcies of First Brands and Tricolor, raising investor concerns about credit quality and fraud risk. Jamie Dimon, CEO of JPMorgan Chase, captured market sentiment by warning, “When you see one cockroach, there are probably more,” which fueled speculation about hidden vulnerabilities in credit markets. The private credit market has grown rapidly, attracting increased scrutiny as the credit cycle matures. These high-profile defaults have prompted questions about whether these events are isolated or indicative of broader risks, particularly in private credit markets, which are inherently more opaque than public markets.

In our view, both First Brands and Tricolor failed due to company-specific frauds rather than broader macroeconomic challenges or poor lending practices indicative of systemic issues. Tricolor operated in a high-risk segment, focusing on subprime auto lending—often to undocumented borrowers—and is alleged to have double-pledged loans across multiple credit lines. First Brands was an aggressive acquirer in the aftermarket auto parts industry, relying heavily on off-balance sheet financing that was poorly disclosed to investors. The company was also accused of double-pledging assets in its supply chain and inventory finance arrangements. Supply chain finance has historically been susceptible to fraud, given the high velocity of relatively small transactions. The fact that both frauds have come to light in a short time frame may reflect late-cycle dynamics, but ultimately they are unrelated events resulting from the actions of a few bad actors.

Importantly, the First Brands and Tricolor frauds were not unique to private credit; they affected a range of investment vehicles, including public asset-backed securities (ABS), broadly syndicated loans (BSL), and large bank warehouse lines. Traditional credit market participants—banks, auditors, and ratings agencies—were exposed and failed to detect the frauds. For example, JPMorgan is facing significant losses from Tricolor, and many collateralized loan obligations (CLOs) are facing losses from First Brands’ BSL. In contrast, most high-quality private credit managers identified warning signs early—such as abnormally high margins, opaque off-balance sheet financings, and management credibility—and largely avoided both situations. The ability of private credit managers to conduct deep, ongoing diligence and maintain close relationships with borrowers provided a clear advantage in risk detection and avoidance. Strong alignment of lenders and the ability to properly conduct due diligence is more important than the specific market segment (public or private) when it comes to avoiding fraud and credit losses.

Private credit markets continue to show solid fundamentals. The Federal Reserve has been cutting interest rates and is expected to make three additional 25-basis point reductions by the end of 2026, which will reduce interest expense for middle-market companies and help alleviate cash flow pressures. According to Kroll Bond Rating Agency (KBRA), the weighted-average interest coverage ratio across middle market loans was 2.3 at the end of third quarter 2025, up from 2.0 a year ago. While loan documentation has weakened in middle-market direct lending—particularly in the upper-middle market segment—the sector has not seen widespread use of liability management exercises (LMEs) that have become common in the BSL market. As a result, default rates in the middle market are expected to remain lower than in the BSL market. As of October 2025, KBRA is forecasting a default rate by volume of 1.5% for the direct lending market in 2025, down from 1.8% in 2024. For comparison, this year’s BSL market default rate may reach 3.8%.

Looking ahead, we believe private credit continues to offer attractive risk-adjusted returns and diversification benefits. In particular, we favor commitments to asset-based finance (ABF) funds in 2026 due to the higher barriers to entry and generally stronger lender protections associated with these strategies. However, as the recent frauds have demonstrated, ABF’s additional complexity is both an opportunity and a risk, requiring more robust due diligence.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.

The post Do the Recent Bankruptcies of First Brands and Tricolor Suggest Trouble Ahead in Private Credit? appeared first on Cambridge Associates.

Investment Implications Through The Tech Cycle

To navigate the AI investment landscape, it is helpful to segment the market into five archetypes that capture the diverse ways in which companies interact with AI:

1. Creators are the pioneers at the frontier of AI innovation—companies developing foundational models, advanced algorithms, the software development toolchain, and specialized hardware that form the core of the technology.
2. Disruptors create a transformative change that goes beyond integrating technology into an existing process, launching new business models that were unimaginable prior to the technological leap (e.g., Uber or Amazon of the internet era).
3. Enablers provide the essential physical infrastructure that makes AI possible, including semiconductors, data centers, and energy solutions.
4. Adaptors are businesses that integrate AI into their operations, harnessing its power to drive efficiency, unlock new business models, expand their market share, and maintain competitive advantage.
5. Finally, the Disrupted are incumbents whose market share or relevance is threatened by the rise of AI-powered competitors.

Each of these archetypes presents distinct investment opportunities and risks across asset classes.

As discussed in Part 1, where we highlighted past technology cycles, this framework echoes the dynamics of the internet era that launched the information age. During that period, Apple, Google, and Microsoft were among the creators, building the platforms and software that defined the new economy. Amazon emerged as a disrupter, fundamentally changing the retail landscape. With the emergence of cloud computing, software-defined infrastructure was developed to manage or enable compute, storage, and networking through software. Companies like Intel and Cisco served as enablers, providing the chips and networking equipment that powered the digital revolution. Today, as AI ushers in another wave of transformation, understanding where companies sit within this cycle is essential for identifying both risks and opportunities across the investment landscape.

Creators and Disruptors

Venture capital (VC) remains a crucial funding source for innovative start-ups engaged in high-risk research and product development. This dynamic drove previous technology waves, such as the internet, mobile, and cloud computing. However, the AI era presents a different landscape. Unlike the cloud era—where established companies were slow to adapt and start-ups captured early gains—many incumbents are now early AI leaders. These companies are cloud-native and deeply integrated into corporate systems. They leverage their scale and distribution to build AI capabilities internally or accelerate innovation by acquiring or investing in VC-backed AI start-ups. Notable examples include Google’s acquisition of DeepMind (which powered Google Brain and Gemini), Microsoft’s early partnership with OpenAI, and Amazon’s partnership with Anthropic. Hyperscalers’ capital expenditures have been extraordinary and are expected to continue as AI technology advances. Key areas of VC investment include large language models (LLMs), supporting software infrastructure, and “applied AI” applications built on this foundation.

As outlined in Part 2, VC investment in AI has reached record highs, with intense enthusiasm and abundant capital pursuing a limited number of high-quality start-ups. Adoption rates have surged across many companies (see Part 1), but much of the early revenue is “experimental,” reflecting trial phases rather than sustainable businesses. This momentum has spurred a wave of new company formations and AI strategy announcements, creating significant “AI noise” in the market. Interest is also growing in “physical AI,” where AI intersects with industries such as manufacturing, construction, healthcare, and aerospace and defense. However, all this frenzy has led to inflated valuations, intense competition, and overfunded segments given its relative infancy. Although AI-first companies have seen rapid revenue growth, its durability is uncertain due to the experimental nature of adoption and the lack of strong competitive moats—even companies with $50 million–$100 million in revenue can be overtaken whereas in prior cycles that typically signaled victory. While a few leaders have already created significant value, many AI start-ups are likely to fail due to oversaturation, poor management, and rapid sector evolution.

Historically, major technology shifts often result in commoditization, and it is rarely clear at the onset which companies will ultimately succeed. The winners are typically those that either build on existing technology through innovation or leapfrog older products and services entirely. For instance, Dell Technologies initially dominated the PC market, EMC led in on-premises enterprise data storage before the transition to cloud solutions, and Cisco was the leader in network hardware before the rise of software-defined networking. AI is likely to follow similar patterns, with rapid change and innovation making it difficult to identify long-term leaders. As open-source competition and verticalized alternatives have driven SaaS commoditization, so too will these forces and the broader open-source community drive further innovation and disruption in AI. Despite these uncertainties, we expect long-term VC returns in AI to remain attractive.

Who will be the winning investors? We recommend diversifying across the AI value chain and managing risks through careful position sizing. Investors should prioritize general partners (GPs) with deep sector expertise, particularly at the foundational and network infrastructure levels, and a proven track record of business building. This expertise—whether within specialist or generalist firms—enables better deal flow, talent identification, and assessment of technical merit. Select specialists for investments where technology risk is high, and generalists for broader investment strategies, leveraging the strengths of both. As AI becomes more widespread and many start-ups incorporate it into their products, investment decisions will increasingly focus on how AI is applied rather than on the technology itself. This trend mirrors previous technology cycles, where, as markets matured, investment success depended more on careful selection and curation than on technical expertise. Many GPs focused on AI are relatively new and still gaining investment experience, given the technology’s rapid rise in prominence. Large generalist firms have captured many early AI successes, often partnering with specialists to combine strengths. These generalists offer larger capital pools, enabling them to support AI start-ups through multiple funding rounds, provide customer access, and offer business-building expertise. Their broad go-to-market and business development capabilities help start-ups as they scale.

Enablers

Enablers are the backbone of the AI revolution, providing the physical infrastructure that supports AI’s rapid expansion. The primary beneficiaries to date have been semiconductor manufacturers (especially those producing AI chips), hyperscale data center operators, and the power and utility companies that support this ecosystem. However, the scale and speed of investment in these areas have raised concerns about sustainability, valuations, and the risk of overbuilding—reminiscent of the internet era’s fiber optic boom and bust.

The rise of generative AI and LLMs has driven unprecedented demand for high-performance chips, particularly GPUs and custom AI accelerators. Companies like Nvidia, AMD, and emerging players such as Cerebras have seen orders and backlogs soar. Supply constraints and technological leadership have enabled leading chipmakers to command premium pricing and margins. Dominant players, especially Nvidia (through its CUDA platform), are building integrated hardware-software ecosystems, creating high switching costs and network effects, but also raising antitrust concerns. Valuations remain high, with Nvidia trading at a forward price-to-earnings (P/E) ratio of 32.3, as of June 30, 2025. While this is below 2024 peaks, it remains vulnerable to correction if AI adoption slows, or competition intensifies. As such, consider modest tilts away from expensive public equity mega-cap tech stocks to reduce valuation risk and enhance portfolio diversification.

Data centers are also major beneficiaries, driven by AI, ongoing cloud adoption, and rising data usage. McKinsey estimates data center capacity demand will grow at an annual rate of about 20% through 2030, with generative AI data centers accounting for a small, but growing share of new demand. Investors should partner with infrastructure and real estate managers with specialized development and operating expertise that are well-positioned to benefit from this supply/demand imbalance. However, transaction multiples have risen materially, averaging 25x EBITDA over the last four years according to Infralogic, compared to a 13.5x average for private infrastructure more broadly. This makes careful underwriting essential for attractive returns. Like other AI infrastructure assets, data centers face risk of overbuilding, as well as regulatory and environmental concerns and constraints such as local opposition and permitting delays. These risks can be mitigated by focusing on managers who can develop assets at lower multiples (e.g., low double-digit EBITDA) and sell into a strong market, often with long-term contracts from investment-grade hyperscalers (e.g., Microsoft, Amazon) seeking development partners. In contrast, speculative and remotely located data centers with more limited utility face heightened risks. From a portfolio construction perspective, data centers offer lower expected returns than private investments in innovative AI firms but can provide returns competitive with broad equities (e.g., 15%–20% target gross IRR) with diversification benefits.

Other enablers, such as utilities and grid infrastructure, have also seen increased demand and capital inflows driven by electrification and digitization trends. McKinsey expects global data center capacity demand between 2025 and 2030 to drive investment in power (including generation and transmission) to total between $200 billion (constrained momentum) to $600 billion dollars (accelerated demand), with $300 billion as their baseline for continued momentum. US on-grid electricity demand is expected to increase 2%–3% per year through 2030 up from virtually flat growth over the last decade, with faster growth in Asia (from a lower base) and slower growth in Europe. While difficult to estimate, rapid AI adoption and potential onshoring in the United States could further boost energy demand. Although AI energy efficiency is expected to improve, associated cost reductions may spur broader adoption, likely resulting in net energy demand growth. Investment in essential electricity infrastructure with inelastic demand is critical. Data centers require reliable power, necessitating redundant infrastructure such as back-up generators and batteries.

All enabler segments have strong growth potential, with chips and data centers experiencing the fastest expansion, but they also trade at heightened valuations and have the greatest exposure to overbuilding. Scale, technological edge, strong customer relationships, and specialized expertise are critical for managing these risks.

Adaptors and the Disrupted

Building on the productivity themes from Part 2, growth equity and private equity-backed companies are increasingly using AI to boost revenue and improve margins. As private entities, they have more flexibility to integrate and scale AI across operations, though successful implementation requires careful execution. While many companies are still experimenting, some are already seeing early benefits in product enhancements and margin gains.

Private equity investors are actively assessing both the opportunities and risks AI brings to their portfolio companies and industries. They look for cost savings through automation (e.g., customer support, onboarding, coding) and revenue growth from AI-driven products (e.g., sales planning, demand forecasting). At the same time, they remain cautious about risks, such as commoditization (e.g., graphic design, digital marketing) and increased competition from low-cost automation (e.g., auditing, document preparation, call centers). Technology-focused managers have an edge due to sector expertise, but both specialist and generalist firms are hiring AI talent to support investment teams and portfolios. The full impact of AI will unfold over time as new use cases and broader adoption and understanding of AI technologies and their impact continue to emerge.

Similarly, public companies must adapt to AI or risk disruption. Investors should focus on active management to distinguish winners from losers and to assess price risk, selecting managers with deep sector expertise. Employ long/short and fundamental strategies to manage risk and exploit valuation dislocations. Public investors face the challenge of avoiding overvalued AI leaders while not overlooking lower-priced companies that may lag behind. Many leading public companies are cloud-native and well-positioned for AI, but investors should consider the entire spectrum of innovators and disruptors. Public market valuations for AI-enabled companies have dropped from their late 2021 peak; forward P/E ratios relative to the S&P 500 Index hit a nine-year low earlier this year, and have since rebounded, but remain below recent historical spikes. This environment favors long/short managers that can identify mispriced companies amid the current AI hype.

As outlined in Part 1, we recognize that non-technological factors—particularly regulatory and policy uncertainty—are increasingly shaping both the AI investment landscape and broader societal outcomes. The concept of Responsible AI (RAI) is gaining more attention as generative AI models and systems grow in complexity and become more deeply embedded across industries. RAI frameworks address the development and deployment of LLMs and broader AI applications, emphasizing principles such as fairness, transparency and explainability, accountability, privacy, safety, and security. From an investment perspective, effective governance is inherently complex, intersecting regulatory, ethical, technological, and human considerations. This complexity necessitates cross-disciplinary collaboration and often involves navigating trade-offs and misaligned incentives. As AI adoption accelerates, reported incidents of ethical misuse have increased in recent years. A recent survey found that only 14% of businesses have dedicated AI governance roles, yet 42% reported improved operations and 34% noted increased customer trust due to RAI policies and investments. ² Companies should proactively assess, and address financially material risks associated with neglecting RAI practices, such as regulatory actions or erosion of their societal license to operate, which could result in negative commercial consequences. Governments worldwide are trying to address complex issues like data privacy, algorithmic transparency, antitrust, and national security, and new regulations could significantly impact sector competition. Investors must also monitor regulatory developments closely, as evolving rules and policies will likely influence long-term value creation and competitive differentiation in the rapidly evolving AI sector.

The “AI noise” phenomenon extends beyond private investments. Most technology companies now market themselves as AI-focused, and those that do not, risk appearing outdated. Enterprise software incumbents with high switching costs, complex technology, and strong innovation pipelines may continue to thrive, while agile start-ups can exploit weaknesses and expand from niche solutions into strategic adjacencies, potentially displacing incumbents. For example, it is unclear whether established security firms will lead in AI security or whether nimble start-ups will secure the AI/ML software supply chain. ServiceNow, a leading enterprise software provider, has thus far demonstrated successful AI adoption by leveraging its integrated suite and existing customer base to pivot toward AI-driven solutions. Given the rapid pace of change, both long-only and long/short hedge funds can find alpha by capitalizing on short-term disruptions and mispriced companies. Valuation-based and fundamental short strategies remain relevant, though it can be difficult to short declining businesses that retain temporary relevance or to identify companies prematurely dismissed as AI losers. Investors should consider managers with crossover expertise—spanning both public and private markets—as they are well-positioned to capitalize on rapidly evolving AI developments by spotting trends in private markets before they are reflected in public market valuations, and can continue to invest post IPO.

AI-related risks and opportunities are increasingly influencing credit markets. Credit managers are financing core infrastructure—such as GPUs, data centers, and energy projects—while also supporting the broader AI ecosystem. Several large managers are establishing dedicated asset-backed finance teams and raising capital specifically to pursue these opportunities. Direct lenders, in particular, have significant exposure to technology and business services, which will need to adapt in response to AI advancements.

More broadly, credit managers must evaluate the adaptability of their portfolio holdings. Many software companies—particularly those with high leverage and business models vulnerable to AI automation (e.g., HR, legal, accounting, and other back-office SaaS providers)—face considerable disruption risk. The past decade’s low-rate environment led to aggressive leverage and high valuations, leaving some companies with thin interest coverage and little margin for error. These firms are especially vulnerable if AI-driven disruption erodes their revenue base. Should AI agents automate or disintermediate core functions, revenue models may be cannibalized, and even modest declines in topline revenue could threaten debt service capacity.

Some credit managers are proactively encouraging portfolio companies to adopt AI, aiming to drive efficiencies and mitigate disruption risk. Lenders are increasingly evaluating management’s AI strategy as part of their underwriting process. Companies that successfully integrate AI may improve margins and creditworthiness, while laggards risk being left behind. As disruption accelerates, a wave of distressed opportunities may emerge among over-levered incumbents unable to adapt to AI-driven change. However, the timing of this transition is highly uncertain: some companies may be “slow melting ice cubes,” experiencing gradual market decline, while others may yet adapt successfully.

Investors should select credit managers who proactively assess AI-related opportunities and risks, including overbuilding in data centers and other infrastructure, while proactively managing exposure to incumbents in sectors vulnerable to AI disruption, such as highly leveraged back-office SaaS providers. Credit opportunity managers may be best positioned to benefit from distressed cycles arising from AI-driven disruption, as these managers can capitalize on market dislocations.

Investors should question managers on their approach to AI, both in terms of portfolio company adaptation and exposure to AI-related risks and opportunities, as part of ongoing due diligence.

Conclusion

AI is fundamentally reshaping the investment landscape, presenting both extraordinary opportunities and new risks across asset classes. The technology’s reach extends from the innovators building core capabilities, to the enablers providing critical infrastructure, to the adaptors and disrupted incumbents navigating a rapidly changing environment. Although substantial investment has already driven rapid growth in AI and its supporting infrastructure, we remain in the early stages of this technological shift, which is expected to evolve over the next decade and beyond. In previous technology cycles, the initial investments and returns from foundational innovation were ultimately surpassed by the gains generated by disruptive companies. These disruptors leverage the established or rebuilt technology infrastructure and benefit from network effects as commercial adoption accelerates, enabling them to redefine industries or create entirely new markets and business models. Attractively valued companies that can leverage AI to improve their profitability should also benefit meaningfully.

Investors should strategically seek opportunities to incorporate AI Creators, Disruptors, Enablers, and Adaptors within their portfolios, all the while maintaining a careful watch on potential disruption risks and the possibility of inflated valuations and overbuilding. Investment success in this new era will require investors to combine deep sector expertise, rigorous due diligence, and a willingness to adapt as the technology and its applications evolve. Investors that partner with managers that can distinguish between hype and enduring value, anticipate regulatory shifts, and identify the true drivers of sustainable growth will be best positioned to capture the far-reaching potential of AI in shaping asset allocation for years to come.

Grayson Kirk, Graham Landrith, and Archie Levis also contributed to this publication.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.
2. Nestor Maslej, Loredana Fattorini, Raymond Perrault, Yolanda Gil, Vanessa Parli, Njenga Kariuki, Emily Capstick, Anka Reuel, Erik Brynjolfsson, John Etchemendy, Katrina Ligett, Terah Lyons, James Manyika, Juan Carlos Niebles, Yoav Shoham, Russell Wald, Tobi Walsh, Armin Hamrah, Lapo Santarlasci, Julia Betts Lotufo, Alexandra Rome, Andrew Shi, Sukrut Oak. “The AI Index 2025 Annual Report,” AI Index Steering Committee, Institute for Human-Centered AI, Stanford University, Stanford, CA, April 2025 and McKinsey & Company survey 2024.

The post Navigating the AI Revolution: AI’s Far Reach in Shaping Asset Allocation Opportunities appeared first on Cambridge Associates.

Heading into 2025, historically low spreads on some higher-yielding credit instruments meant investors were not well positioned for recent tariff-related turbulence. At the end of 2024, the option-adjusted spread (OAS) on US HY bonds stood at 287 basis points (bps), in the bottom decile of historical observations. As a result, while the backup in spreads in recent weeks felt dramatic, it still leaves the current OAS (409 bps) below its historical median. A similar trend is evident in loans. The discount margin on BB-rated loans has widened by approximately 40 bps in 2025, but the current 297-bp spread is only around the 45th percentile of historical observations.

Investors considering increasing allocations to these assets should recognize spreads could go significantly higher if the economy enters recession. While the Global Financial Crisis may be an extreme level for comparison (HY spreads reached nearly 2,000 bps), during the past three recessions HY spreads averaged around 800 bps, around double today’s level. Another consideration is whether current pricing suggests HY bonds can keep pace with a comparable stock/bond mix. Our data suggest that buying HY bonds around current spreads (in the second quartile) has often resulted in underperformance relative to a stock/bond mix. Conversely, HY bonds have typically outperformed when spreads rise to the top quartile (around 585 bps or higher). Investors may be better off waiting for spreads to reach these higher levels before increasing allocations.

While the macro environment remains uncertain, there are positive arguments to be made in favor of US HY bonds and loans. Entering what may be a period of subdued growth, many HY issuers are in a position of relative strength. Rising revenue and earnings have allowed companies to gradually deleverage in recent years, and metrics like interest coverage ratios have shown steady improvement. Notably, today’s HY index consists of higher-quality borrowers than historically has been the case, which could provide a buffer if conditions worsen. Currently, ~53% of the HY index carries at least one BB rating, an 8 percentage point increase from a decade ago.

HY bonds and loans may also benefit from investors attracted to their higher coupons. The current HY bond index yield of 8.4% is around 170 bps above its average over the past decade. While broadly syndicated loan yields—currently around 9.0%—may also look enticing, we caution that this reflects lower average credit quality. Additionally, these instruments may see coupons decline if, as expected, the Federal Reserve resumes its rate-cutting cycle in 2025.

Given the uncertainty surrounding tariff-related volatility and concerns over foreign demand for US assets, investors should hold off on adding HY and loan exposure. Also, certain pockets within liquid credit already appear more attractive. One example is CLO mezzanine debt, which currently offers a discount margin of around 775 bps (equivalent to around a 11.5% yield). Historically, this asset class has suffered lower defaults than comparably rated HY bonds, though its lower liquidity can result in higher mark-to-market volatility. Due to the dispersion in underlying CLO fundamentals, we believe this asset class is best accessed via skilled managers.

In summary, HY bonds and loans have sold off in recent weeks, but from historically expensive levels. We recommend waiting for further clarity on the macro outlook or further pricing improvements before adding exposure to assets like HY bonds and loans. When conditions improve, investors contemplating adding to credit allocations should also consider CLO debt, which is currently more reasonably priced but may still face spread widening if tariff-related volatility escalates. Meanwhile, investors should maintain allocations to high-quality sovereign bonds, which should continue to provide critical portfolio diversification and stability amid ongoing macro uncertainty.

Footnotes

1. The situation is somewhat flipped in China, where access to the most advanced chips remains a constraint, while electricity is more available. Notably, China controls refinement of most critical material, such as rare earths.
2. Nestor Maslej, Loredana Fattorini, Raymond Perrault, Yolanda Gil, Vanessa Parli, Njenga Kariuki, Emily Capstick, Anka Reuel, Erik Brynjolfsson, John Etchemendy, Katrina Ligett, Terah Lyons, James Manyika, Juan Carlos Niebles, Yoav Shoham, Russell Wald, Tobi Walsh, Armin Hamrah, Lapo Santarlasci, Julia Betts Lotufo, Alexandra Rome, Andrew Shi, Sukrut Oak. “The AI Index 2025 Annual Report,” AI Index Steering Committee, Institute for Human-Centered AI, Stanford University, Stanford, CA, April 2025 and McKinsey & Company survey 2024.

The post Should Investors Add to High-Yielding Credit Allocations, Given the Recent Rise in Spreads? appeared first on Cambridge Associates.