How diversified is your AI exposure if several parts of the portfolio ultimately depend on the same underlying thesis?
A portfolio can appear diversified across public equities, venture capital, growth equity and infrastructure while still carrying significant exposure to the same AI value chain. Understanding that overlap requires looking beyond asset-class labels to the economic drivers underneath them.
A family office may see AI exposure spread across several parts of the portfolio. Public equities sit in one report. Venture and growth investments sit in another. Infrastructure may be reviewed separately again.
On paper, those positions can look diversified. In practice, they may still depend on the same underlying AI spending cycle, valuation assumptions, and demand from a relatively small group of companies.
That is the portfolio problem this article examines.
Goldman Sachs’ reports that a majority of global family offices plan to overweight their portfolios to the technology sector over the coming year. This represents the single largest sector tilt and compares to only 5% that are maintaining an underweight position. BNY Wealth’ reports that AI is a top conviction theme for 83% of single-family offices over a five-year horizon. J.P. Morgan Private Bank’s Global Family Office Report confirms that 65% cite AI as their primary thematic priority. The question then becomes – how this capital is deployed and what issues does this create.
The vast majority of this capital remains heavily concentrated within a relatively small range of mega-cap public equities and late-stage private company growth rounds. The issue is whether family offices are aware and prepared for the systemic vulnerability that has been created.
Structural Diversification?
Many family offices do not limit their exposure to listed and unlisted mega-cap technology stocks. Family offices also allocate to venture capital funds chasing outsized returns from early-stage AI companies. In addition, offices are seeking exposure to the sector via deploying capital directly into digital infrastructure and seeking longer duration more resilient cash flows (eg data centres, power grid developments). On its surface, this presents as classic diversification – Public Equities, Venture Capital & Growth Equity, and Real Assets & Infrastructure.
The issue, however, is the lack of traditional correlation that exists between supposedly diversified exposures and the large extent to which these are different exposures to that same thesis. These risks are exacerbated by the following factors.
- Private Capital Concentration: Private funding is heavily weighted to a small cohort. Recent increases in AI valuations may have removed a lot of the asymmetric upside and a downward repricing can reprice the entire basket.
- Infrastructure Interdependence: Digital infrastructure relies heavily on persistent demand and capex from the same small cohort of AI firms. If capex contracts or model monetization slows, the infrastructure layer faces utilization and counterparty risks.
- Correlation Convergence Under Stress: All three sleeves rely on the continuous monetization of a single technological shift. In a broader thematic drawdown or prolonged correction, cross-asset correlation increases dramatically.
Practical Insights for Portfolio Governance
Market commentators and institutional advisors recommend the following practical steps:
- Conduct Look-Through Portfolio Audits: Quantify other hidden exposures within ETFs, VC, PE and direct investments to establish true aggregate concentration.
- Consistent Review and Analysis: Regularly review the aggregate exposures to AI related investments and rebalance accordingly, and conduct portfolio stress testing an "AI Winter" scenario.
- Institutionalize Due Diligence: Ensure effective decision-making via access to internal and external technology expertise to properly assess risk and return.
- Implement Cross-Sleeve "Look-Through" Exposure Limits: Establish single-theme thematic concentration limits that operate across the entire balance sheet.
- Rebalance Asset-Classes: Review and implement appropriate levels of true non-correlated exposure by balancing AI themed investments with uncorrelated investments.
For investment teams, the first difficulty is often assembling the information required to calculate the combined exposure.
Public securities, fund holdings, direct investments, infrastructure positions and supporting documents may sit in different systems. Altius brings portfolio data and documents into a common context, allowing teams to examine risk, concentration and underlying relationships across the portfolio rather than one sleeve at a time.
