AI concentration: a new diversification challenge for investors
Artificial intelligence has the potential to be one of the most transformative technologies in human history. The rapid advancement of large language models and increasing adoption of AI across businesses has created substantial opportunities for investors, while also raising hopes of productivity gains, stronger economic growth and improved living standards over the long term.
However, alongside these opportunities, investors may need to consider an increasingly important question. How much AI exposure do they already have?
The issue is often framed as one of equity market concentration, particularly in the US technology sector, but the reality is broader. AI-related companies now represent a substantial proportion of major equity indices, AI infrastructure spending has become an important driver of economic growth and the financing of that infrastructure is increasingly influencing corporate debt markets. As a result, a theme that initially appeared confined to a handful of technology companies is becoming embedded across multiple asset classes and economic activities.
AI has become increasingly difficult to avoid
The most obvious area where AI concentration is visible is equity markets. The largest company beneficiaries of the AI theme have generated extraordinary returns over recent years, leading to significant increases in their index weights. Technology companies remain at the centre of the investment story, but the impact of AI is no longer confined to the technology sector.
The build out of AI infrastructure requires semiconductors, data centres, power generation, cooling systems, specialist materials and increasingly complex supply chains. As a result, sectors such as industrials, materials, utilities, energy and financials have also benefited from the investment cycle. At the same time, healthcare companies, manufacturers and other businesses are exploring how AI can improve productivity and create new products and services.
This broadening of the opportunity set is positive for markets and economies. Strong earnings growth across both the hyperscalers and wider corporate sector has helped distinguish the current environment from previous technology booms where valuations often raced ahead of underlying profitability. However, it also makes the theme increasingly difficult to avoid. Investors who believe they are diversified across sectors, regions and industries may still be exposed to the same underlying driver.
The challenge is not limited to the United States. AI-related companies represent a growing share of developed market indices, while many emerging markets have become increasingly tied to the supply chain underpinning AI infrastructure, particularly through semiconductor manufacturing, hardware production and digital platform businesses. The result is that AI exposure has become a global phenomenon rather than a purely US technology story.
This also raises questions around passive investing. Historically, broad market index funds provided exposure across hundreds of companies and a wide range of economic drivers. Today, market-cap weighted indices naturally allocate increasing amounts of capital towards the largest winners. Investors remain diversified from a stock-count perspective, but arguably less diversified from a thematic perspective than they were fifteen or twenty years ago.
The spread into other asset classes
AI concentration is no longer solely an equity market issue. The extraordinary scale of investment required to build new data centres and computing capacity is increasingly being financed through debt markets. There has been a significant increase in borrowing by major hyperscalers and AI-related businesses to support infrastructure expansion. Corporate bond investors therefore may have more exposure to the AI investment cycle than is immediately obvious.
More broadly, AI has become an increasingly important contributor to economic growth through spending on software, research and development, information processing equipment and data centre construction. AI has recently been a significant contributor to US economic growth through the substantial investment associated with the AI infrastructure build-out, while strong equity market returns have further supported household wealth and consumer confidence.
Taken together, this means that equity markets, credit markets and economic growth are becoming increasingly influenced by the same underlying investment theme. This does not mean AI is a bubble, nor does it imply that the investment case is weakening. It simply suggests that diversification across asset classes may not provide quite the same degree of risk reduction as investors have historically assumed.
What could go wrong?
The central case remains compelling. AI adoption continues to accelerate, earnings growth remains strong and the potential applications across industries appear vast. It is entirely possible that AI proves to be one of the defining technological developments of the century.
Nevertheless, concentration risk exists regardless of how attractive a theme appears.
Investors do not need AI to fail completely in order for market leadership to change. A period of slower monetisation, higher financing costs, increased regulation, energy constraints, geopolitical tensions or stronger competition from China could reduce enthusiasm towards AI-related assets. Equally, a broader economic slowdown could cause businesses to moderate capital expenditure plans and delay investment in new infrastructure.
History demonstrates that markets tend to underestimate risks when confidence is highest and a dominant investment narrative becomes widely accepted.
How can investors diversify?
For most investors, the answer is unlikely to be avoiding AI altogether. In fact, betting against transformational technological progress has historically been a difficult strategy to execute successfully. Given the potential economic benefits and earnings opportunities on offer, maintaining some exposure to the theme is likely to remain important.
Instead, the more relevant question is whether portfolios contain sufficient diversification should sentiment towards AI weaken.
Factor investing may offer one potential solution. During the AI-led market weakness seen in July 2026, factors such as value and minimum volatility proved more resilient than growth and momentum-oriented strategies. While one month does not establish a trend, it provides a useful illustration of how different factor exposures can behave when leadership rotates.
Value strategies typically focus on companies trading at lower valuations and often have lower exposure to highly rated growth businesses. Minimum volatility approaches favour companies with more stable earnings and share price characteristics. Both can provide diversification away from the areas of the market most closely associated with AI enthusiasm.
Beyond equities, investors may also consider traditional diversifiers such as high-quality government bonds and gold. In an environment where AI-related assets experienced a meaningful repricing, a flight to quality could potentially support these areas of the market. Alternative strategies with lower dependence on equity market returns may also help reduce overall portfolio sensitivity to a single dominant theme.
Another potential approach to gain diversification is to seek out investment strategies not tied to broad market performance, including manager skill based strategies such as hedge funds. Investment in physical assets, such as timberland, agriculture and even more traditional real estate, could also have a role to play.
The goal is not to avoid AI. Rather, it is to ensure that portfolios are not solely dependent upon it.
Artificial intelligence may ultimately justify the optimism surrounding it. The productivity gains, economic benefits and commercial opportunities could be substantial, and investors who ignore the theme entirely risk missing one of the most important investment opportunities of their careers.
Yet successful investing has never been solely about identifying the winners. It has also been about managing the consequences of being wrong.
AI's growing influence across equities, credit markets and economic activity means many investors may already have greater exposure to the theme than they realise. As the opportunity grows, so too does the importance of asking whether portfolios remain sufficiently diversified. Which in itself, is a call to at least examine the alternatives.
John Plenderleith, Investment Consultant
Click here to find out more about John Plenderleith