Anthropic heads for a listing: how much artificial intelligence is already in portfolios?

A global equity fund and a bond fund can perform different functions and, at the same time, have something in common: part of their performance may depend on how large technology companies finance and make productive their investments in artificial intelligence. Anthropic’s expected listing, which has applied for admission to Nasdaq, offers an opportunity to observe this overlap. Those who do not subscribe to the IPO and do not intend to buy the stock may already have a part of the economic risk that the debut will bring under the spotlight in their portfolio.

This is the central point of the note by Jordy Hermanns, strategist and portfolio manager, and Jacob Vijverberg, head of asset allocation, both at Aegon Asset Management. Their analysis starts from the valuation of a single company and reaches portfolio diversification. Owning hundreds of stocks reduces the risk linked to the fortunes of a single business. Yet the question remains how many of those stocks respond to the same factors: spending on data centres, the cost of financing, the ability to sell artificial intelligence services at prices sufficient to remunerate the investments. It is a question that concerns both equities and corporate credit.

A daily price for AI expectations

A diversified technology group can support investments in artificial intelligence with revenue from advertising, software or other mature activities. In the case of Anthropic, the industrial case is more directly linked to the development and sale of artificial intelligence models. For Hermanns and Vijverberg, the listing will therefore make investors’ judgement more visible on a basic question: will the revenue generated by the models be sufficient to justify the spending on training them, running them and serving a growing number of users? The share price will offer a daily signal on expectations; what gives that signal substance will be margins, cash and the cost of computing capacity.

Anthropic’s valuation also matters to technology groups that hold stakes in the company. The authors note that, for some groups, gains on stakes have already contributed to accounting results. A decline could produce the opposite effect, depending on the nature of the stake held and the applicable accounting treatment. The point, for anyone reading a balance sheet, is to distinguish the growth in profits generated by the core business from the effects of revaluations. A change in Anthropic’s price can affect the accounting results of anyone holding a stake in it without changing, at the same time and to the same extent, that group’s sales or margins.

More companies in the index do not necessarily mean more diversification

The listing and entry into a global index are two distinct events. Admission depends on the rules of the individual index provider; weight is also determined by the free float, that is the portion of shares available to public investors, and by price. A very large market capitalisation therefore does not automatically translate into an equivalent weight in the portfolio of an index fund. But when a new stock is admitted, the index funds that replicate that basket adjust their exposure in line with the methodology applied.

Aegon Asset Management examined this effect on the MSCI ACWI, the global index that includes developed and emerging markets. In the illustrative scenario constructed by the authors, several large listings, considered together and with assumptions on free float, could push the weight of the United States towards 65% and increase exposure to technology. That number is not a forecast or an estimate of the effect of Anthropic alone. It serves to show how a global index may add companies and, at the same time, become more exposed to the countries and sectors that already have the greatest influence on its performance.

The second door: corporate credit

The most original part of the note concerns corporate bonds. The major cloud operators and other companies linked to artificial intelligence also finance themselves in the corporate credit markets, in dollars and in euros. The capital raised may help support investments in data centres, networks and computing capacity infrastructure. For holders of a bond fund, the link to artificial intelligence can therefore run through the soundness of the issuers financing that spending, even without a direct investment in their equities.

The shareholder and the bondholder remain exposed to different outcomes. For the shareholder, the issue is how much value remains after costs and investments. For the bondholder, what matters is the issuer’s ability to pay interest and repay debt, together with changes in rates and the premium demanded by the market for credit risk. If investments in artificial intelligence generated less cash than expected, the same development could weigh on both equity valuations and corporate credit, with different timing and intensity. This is the risk that can be overlooked when diversification is assessed by looking only at the split between equities and corporate bonds.

Diversification is also measured by risk factors

In the comparison proposed by Aegon Asset Management, Europe has a market for listings on a different scale from that of the United States. The smaller presence of exceptionally large debuts may translate into a different sector composition, without on its own indicating which market will deliver the best returns. And a European stock may be exposed to the same spending on artificial intelligence through its customers, while a bond denominated in euros may finance a US group. Geographic diversification, currency diversification and diversification by risk factors answer different questions.

To read a portfolio through this lens, it is worth starting with the exposures that add up: how much do the United States and technology account for in the global equity fund? Which issuers also recur in the bond fund? And how much do they depend, directly or through customers and suppliers, on the same cycle of investments? Anthropic’s debut will put a public price on one of the emblematic companies of artificial intelligence. For many investors, the most useful question will be how much of that economic cycle they already have in portfolio.

Source of analysis: Jordy Hermanns and Jacob Vijverberg, Aegon Asset Management, «Megacap IPOs in 2026 (2): Anthropic puts the AI boom to the test», November 2026. SEC preliminary prospectus