Portfolio Scenario Analysis: AI Market Correction Impacts on Investments Dissected

The anticipation of Q2 earnings results from the big-tech companies centered on whether skyrocketing spending on AI will be justified by sustainable returns, and it amplified fears of a possible AI market correction. The fears of insufficient return on the investments in AI are coupled with growing concerns about the rising costs to support AI growth. While some of the companies with AI focus reported accelerating revenue related to AI spending, others demonstrated less-clear expectations for longer-term returns.

In this article we review our scenario analysis of the possibility of an AI bubble burst and its impact on financial investments split in standard investment strategies.

An artificial intelligence market correction would unfold sequentially, starting with underwhelming returns on enterprise AI investment, spreading into a sharp decline in tech-company valuations and companies where business is centered around AI. This will then spread across the rest of the equity market, ultimately risking the wider credit market and causing economic strain.

We explore each of these phases of our scenario analysis in more detail below and test their impact on standard investment strategies for asset managers to visualize the impact. Asset managers can also further explore portfolio implications with their own assumptions.

Phase 1: Disappointing enterprise returns for the Magnificent 7 as representatives of the large businesses, driven by AI technology.

In this phase we want to slice the impact on investments in two subsequent effects: the immediate one when only the Magnificent 7 tech companies drop and then the subsequent drag on the rest of the tech- and AI-focused companies.

How do we define which companies are AI-related?

We use a two-fold approach to define which stocks from the representative equity index could be considered AI. We define the AI-focused stocks by adding to the Magnificent 7 the “pick and shovel” companies that supply underlying hardware, computing components, manufacturing tools, power systems, and specialized software required to build and deploy artificial intelligence.

As an additional filter to also capture companies where the business model is not easily defined as “pick and shovel”, we select stocks with high correlation to the core AI stocks. For example, companies where business is highly dependent on AI-focused companies, such as steel producers supplying material for AI data centers or companies recasting their investment narrative around the physical hardware required for artificial intelligence. In that context, we identified ~60 out of 500 companies from a broad large-cap US equity index with AI-related profiles.

How do we define the size of the shock?

For the size of the price drop among the stocks in our analysis, we selected an average of -15%. That number is based on how the market responded to valuations of companies with disappointing Q2 earnings reports, which varied from -6% to -7% and -22% to -25%.

Why both correlated and uncorrelated scenarios?

For the immediate shock we’ve run both correlated and uncorrelated scenarios shocking the Magnificent 7 down and then another uncorrelated scenario shocking the stocks where business is strongly related to AI.

Uncorrelated scenarios apply shock only to Magnificent 7 stocks, with zero return for the rest. That type of shock allows us to measure the drop in Magnificent 7 stocks as isolated vulnerability. We then compare it to the correlated scenario, which shows the systematic impact from a Magnificent 7 stock price correction spreading across other companies in the following days/weeks. We define the correlated scenario by applying a drop of -15% to the Magnificent 7 and then calculate the hypothetical return on the rest of the assets, based on their correlation to the Magnificent 7.

Additionally, we analyze a scenario where all AI-related stocks drop and compare it to the above two scenarios to measure the diversification of the index/portfolio with regard to AI-related shocks.

Below we see the results from the scenarios on a few standard investment strategies as of July 31, 2026.

Several observations:

As expected, investment strategies with a heavy tilt toward equities will be more impacted by these shocks given they are equity-centric at this phase of the crisis. The heavier the tilt, the higher the drop in the return.

Uncorrelated equity scenarios will have no impact at all on pure bond funds, as they do not hold the stressed equity positions. There is some effect for bond assets under the correlated scenario, which is calculated via the correlations of bond risk drivers to the equities stressed under the scenario.

When we compare the impact from the Magnificent 7 drop in the uncorrelated and correlated version of the scenario, we can measure the size of the systematic effect. As of the end of July, the Magnificent 7 have a nearly 30% weight in the US large-cap equity index that we analyzed. The difference in the above chart for the Aggressive Growth strategy (100% equity) shows -7.8% in the correlated scenario and -4.6% in the uncorrelated scenario.

The drop in the Magnificent 7 results is almost equal to the drop for the rest of the index, which indicates that those few large companies drag the rest of the index as well. If the difference in the two shocks was smaller, the risk of the Magnificent 7 stocks would be more idiosyncratic in nature than systematic.

Comparing the AI-related stocks’ uncorrelated shock (-7.1% on a pure equity index) with the Magnificent 7 correlated shock (-7.8%) gives another indication of the AI-driven impact on the index. The difference between the two scenarios is very small; less than 1% drop in the return. In the uncorrelated AI-related stocks shock, we shocked only ~60 stocks down by 15%. They have almost 50% weighting and cause a drop in the index almost the same as when the Magnificent 7 drop by 15% and the rest of the market follows based on their correlations.

Analysis of the uncorrelated and correlated scenarios above serves as a good indication of how diversified an equity portfolio or index is with regards to Magnificent 7 or AI-related risk drivers.

Phase 2: AI companies drag the wider equity market down as market gains are heavily concentrated in a handful of mega-cap tech giants. The forced selling and redemptions pull down the broader stock market.

While some sectors may be less impacted by artificial intelligence market correction, similar to the Dot-Com bubble burst in March 2000, in this hypothetical second phase the tech-heavy Nasdaq index is expected to collapse.

Therefore, a practical way to explore the second phase of impact is to use the correlated scenarios where we shock the Nasdaq 100 down by 35%. This drop replicates the short-term response to the tech market in the Dot-Com bubble burst.

Phase 3: Debt markets follow the market correction.

In this last phase, the impact from an AI market correction is expected to be visible on the bond market as well with the credit spread widening. Although now the companies in debt related to AI are large and stable in terms of business, the depreciated valuation of the technical companies in an AI bubble burst will make it more difficult for them to borrow money and will increase the cost of borrowing.

To replicate this phase, we add to the NASDAQ index a drop of 35% from the phase 2 scenario, a drop on bond markets of 3% (similar to the Dot-Com crisis). We have previously explored this as the short-term response to a potential AI bubble burst scenario in a previous post: Stress Testing Amid Rising Fears of an AI Bubble.

A few observations below for the stress test as of July 31, 2026:

The difference in the magnitude of the impact from the hypothetical AI bubble burst in Phases 1 and 2 is notable. The main explanation being that Phase 2 takes a couple of months, during which the Magnificent 7 company valuations continue to drop and the size of the shock is now -35%, compared to -15% in the first phase.

We can quickly check this by comparing a scenario where we shock only Magnificent 7 by 35% correlated and NASDAQ by 35% correlated. Results below show pretty similar impact. This once again solidifies the observation of the predominant effect Magnificent 7 companies have on the large-cap or tech-tilted indices.

In Phase 3, bond markets also suffer some loss due to higher credit spreads and higher cost of debt. That impacts all investment strategies, though to a lesser magnitude than the market correction from the previous phases. As expected, it adds more stress for strategies with a higher tilt toward fixed-income assets.

Conclusion

AI-related scenarios allow the exploration of a hypothetical AI market correction and the impact that could have on investment portfolios or investable indices. Decomposing the impact in three separate phases draws a more detailed picture of how hypothetically different investment strategies would react to each component of the market shock.

In a hypothetical AI-driven market downturn all equities are impacted, with the magnitude of the impact rising with the correlation of the company to the Magnificent 7 companies. As equity markets are widely driven by the Magnificent 7 companies, the market correction from them will quickly translate to the rest of the equity markets. In such scenarios, investment strategies heavily allocated to fixed income will be safer investments but will suffer at a later stage when the cost of borrowing increases.

In addition, correlated and uncorrelated stress tests serve as a great starting point in analyzing the isolated and systematic impacts that certain shocks may have on portfolios or indices. The comparison between the two may indicate how diversified the portfolio or index is with regard to the shock analyzed.

All of these types of scenarios and shocks are readily available with the enhanced functionalities of FactSet Stress Testing Module 2.0.

This blog post is for informational purposes only. The information contained in this blog post is not legal, tax, or investment advice. FactSet does not endorse or recommend any investments and assumes no liability for any consequence relating directly or indirectly to any action or inaction taken based on the information contained in this article.