Diversified on Paper, Concentrated in Reality: A Closer Look at Portfolio Risk

At first glance, the portfolio looks diversified. Forty stocks. Multiple sectors. Technology, Industrials, Financials, Utilities. On paper, it checks the classic diversification boxes. But when we look more closely to the market’s underlying dynamics, a different story emerges. Many of those holdings, despite their different labels, are marching to the same beat. That beat is artificial intelligence (AI).

As we discussed in a recent TDAM Talks Podcast, diversification today isn’t just about how many stocks you own, it’s about how many independent drivers you truly have. In today’s market, AI has quietly become a dominant force linking businesses that once behaved very differently.

Consider the current equity landscape. AI optimism has propelled a small group of mega‑cap technology companies, the "Magnificent Seven", to represent more than one‑third of the S&P 500’s total market capitalization, up from roughly 20% at the end of 2022¹. That concentration alone has implications for index investors. But the ripple effects don’t stop there.

Look beyond Big Tech:
 

• Industrials are benefiting from demand for data‑centre construction, cooling systems, networking equipment, and advanced manufacturing tools needed to support AI infrastructure²
• Energy and Utilities, traditionally defensive, are now tied to surging electricity demand from power‑hungry data centres³
• Semiconductors and equipment makers are seeing record order backlogs as chipmakers raise capital spending to expand AI‑related capacity⁴
• Financials are financing this build‑out, from data‑centre REITs to the debt funding behind hyperscaler capital expenditures⁵
   

On the surface, these exposures look diversified. In reality, they’re increasingly tethered to the same macro assumption: that AI investment continues at scale.

History shows that market concentration doesn’t just affect returns, it affects correlations. Research highlights how correlations among large technology leaders have risen during periods when a single narrative dominates performance, reducing the benefits of owning many names that react similarly to the same shocks⁶.

This dynamic extends across sectors when a theme is broad enough. Owning a utility stock tied to data‑centre power demand, a semiconductor manufacturer supplying AI chips, and a cloud platform monetizing AI services may look like diversification, but if all three wobble when AI spending slows, the portfolio’s vulnerability is revealed.

This is the false comfort we often reference. Risk isn’t removed, it’s simply repackaged.

This doesn't mean investors should avoid AI. Far from it. AI is reshaping productivity, capital investment, and competitive advantage across the economy. But it does mean diversification requires a more rigorous lens. From our perspective, true diversification asks better questions:
 

• Are return drivers genuinely independent?
• How does a portfolio behave if AI spending slows, shifts, or becomes less profitable?
• Where are risks concentrated? Not by sector, but by theme?

As innovation cycles mature, leadership will eventually broaden. But those transitions can be uneven and uncomfortable, particularly for portfolios that unknowingly leaned too hard on a single narrative.

A fair question naturally follows: Even if diversification is being quietly undermined by powerful themes like AI, are active managers any different?

When a single force reshapes large parts of the economy, avoiding it entirely is neither realistic nor desirable. Many actively managed portfolios, including those we help manage, do have exposure to companies influenced by AI, even when those companies sit across very different sectors. The risk isn’t having thematic exposure. It’s not knowing you have it.

Sector diversification alone doesn’t guarantee independence of outcomes. Holdings may span technology, financials, industrials, or energy, yet still be influenced by the same underlying investment cycle, such as AI‑driven capital spending. The difference is intent.

What active management allows us to do is identify where thematic overlap exists, understand how it shows up across sectors, and size that exposure intentionally. An index may inherit concentration simply because of market capitalization. An active portfolio can acknowledge that concentration, assess whether it is being adequately compensated, and adjust as conditions evolve.

In that sense, the presence of cross‑sector AI exposure in an active portfolio doesn’t contradict concerns about modern diversification, it reinforces them. It underscores why understanding what truly drives returns matters.

This is where we believe active management plays a critical role. Active managers aren’t limited to owning what the index owns or accepting concentration simply because it exists. We can look through sector labels, assess true drivers of risk and return, and make deliberate decisions about where exposure is rewarded and where it may be redundant.

Active management allows for selective participation, maintaining exposure to long‑term structural opportunities like AI, while seeking to manage concentration, valuation risk, and unintended overlaps. Just as importantly, it enables informed judgment during periods of transition, when dispersion widens and fundamentals begin to matter more than headlines.

In an environment defined by fast‑moving themes and uneven outcomes, active management isn’t about predicting the future, it’s about navigating it thoughtfully, with intention, discipline, and a clear understanding of what truly drives results.