Featured Image

One Year Later: A Review of Our Methodology to Analyze the AI Thematic Universe Derived from Supply Chains

Companies and Markets

By Hiroki Miyahara  |  September 11, 2024

During the initial generative AI boom, we presented a new methodology to derive AI companies from supply-chain relationship keywords and analyze their performance rise. The goal was to identify and evaluate potential winners and losers from an investment perspective. Now just over one year later, we are revisiting our same data-driven methodology to analyze the changes.

Performance among AI companies

In the previous article, we analyzed the portfolio performance of AI companies between June 2022 and June 2023. To see how the performance trend has changed, we employed the same techniques to define AI companies.

Figure 1 shows how our sample portfolio of AI companies has performed since June 2023. It consists of 35 global market benchmark constituents that are weighted based on market value.

There has been an upward trend in performance from January 2023 until July 2024. There was also a spike in July 2024 while the performance picked up to some extent afterwards.

Figure 1: AI portfolio performance since June 2023

01-ai-portfolio-performance-since-june-2023

Now, let’s look at the portfolio’s sector allocation and compare the sector performance with the benchmark. Figure 2 shows the average weight and total return by the RBICS L2 sector alongside the benchmark L2 sector returns to determine if the outcome came solely from sector allocation. The return horizon begins on June 30, 2023.

The portfolio still has the largest exposure in Software and Consulting, but the highest performance exhibits in the Electronic Components and Manufacturing sector. There are also mixed results for some other sectors; our analysis reveals both overperformance and underperformance.

However, underperforming sectors have small weights. This implies that filtering companies by only relevant sectors to technology and industry would improve portfolio performance.

Figure 2: Average weight and total return by RBICS L2 sector alongside the benchmark L2 sector returns

02-average-weight-and-total-return-by-rbics-l2-sector-alongside-benchmark-l2-sector-returns

We also analyzed the portfolio performance of AI companies and their suppliers. As of June 2023, there were 204 AI companies and 188 tier 2 suppliers (companies with at least four customers). Among those, 48 are constituents of the global market benchmark.

Figure 3 shows the average weight and total return of tier 2 AI suppliers in each sector of the global market benchmark between June 2023 and July 2024. The result is very similar to the one for the AI company portfolio. The total return still outperforms the global benchmark. In sector performance, the portfolio outperformed in some sectors and underperformed in sectors with very small weights.

Figure 3: Tier 2 AI suppliers portfolio performance

03-tier-2-ai-suppliers-portfolio-performance

Conclusion

In this article, we analyzed the portfolio performance of AI companies derived from supply-chain relationship keywords after the research horizon of the previous article. This analysis helped capture how AI companies' stock prices performed after the initial AI boom. You could replicate this methodology throughout the current growth phase of AI and beyond.

 

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.

Hiroki Miyahara

Principal Product Manager

Mr. Hiroki Miyahara is a Principal Product Manager at FactSet, based on Tokyo, Japan. In this role, he covers the Asia Pacific region for FactSet proprietary content sets including supply chain, RBICS, GeoRev, shipping, and FactSet Data Management Solution. Mr. Miyahara joined FactSet in 2011 and previously held roles as an account executive and product developer. He earned an MSc in economics from the University of Essex.

Comments

The information contained in this article is not 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.