March 29, 2018

Two Ways to Access the Upside in Artificial Intelligence Two Ways to Access the Upside in Artificial Intelligence

By Roberto Lazzarotto, Global Head of Sales, STOXX

This year STOXX Ltd. introduced two indices that give exposure to the growing theme of artificial intelligence (AI).

The STOXX® AI Global Artificial Intelligence Index (AI Global AI index) and the STOXX® Global Artificial Intelligence Index (Global AI index) were launched on the same day in January, and the occurrence was no coincidence. As information technology and autonomous machines develop at fast pace, investors are seeking to profit from AI’s economic impact. Amid this interest, it made sense to us to provide more than one solution.

Their names are very similar, but the composition methodology and economic rationale of the two indices are not. The AI Global AI index picks companies investing in the development and adoption of AI-related intellectual property, via an innovative selection process that relies itself on an AI system. STOXX has partnered up with Yewno, whose knowledge-graph algorithm detects which awarded patents worldwide relate to AI. As such, the STOXX AI Index is the first thematic index to select its constituents by means of AI technology.

The Global AI index, on the other hand, follows the same methodology applied by other STOXX thematic indices. Based on FactSet’s Revere industry hierarchy, the world’s most comprehensive business classification, companies with the highest revenue exposure to AI are selected into the index.

Another important difference is the weighting factor. The AI Global AI index is equally weighted, meaning it allocates the same resources to all members regardless of size. The Global AI index is weighted by the free-float market cap of the selected stocks multiplied by their revenue exposure to AI, with caps to those weightings. Member composition is reviewed quarterly in the former, and annually in the latter.

Geographic and sector concentration

As could be expected, the two approaches result in distinctive index composition and fundamentals. The AI Global AI index, for example, shows higher geographic diversification. While 53% of its constituents are US-based, that ratio jumps to 95% for the Global AI index. Sector-wise, 47% of AI Global AI constituents are technology stocks. In the Global AI index, that sector accounts for 88% of the weighting.

The higher diversification allows the AI Global AI index to track the AI pioneers regardless of their market capitalization. This makes the index a unique platform that looks forward to capture the true winners of future AI growth.

In summary, investors can choose between two rationales: gaining exposure to a wide basket of companies that are spending in AI research and development, or investing in businesses that are already raising the most revenue from AI technologies. The traditionally built index selects companies from focused sectors, while the AI Global AI index looks one step beyond and offers exposure to future value of current investments to reap the benefits of a long-term megatrend.

Different valuation profiles

When it comes to fundamental ratios, both indices also look quite different. The AI Global AI index trades at 15.8 times expected earnings (excluding negative estimates), and 2.5 times its members’ book value.1 Those ratios are higher for the Global AI index: 22.2 and 5.3, respectively.

The latter’s premium may be explained by the strong performance in recent years of heavyweight US stocks, including Alphabet, Facebook, Nvidia and Intel. Those companies account for a third of the index’s weight at the moment.

The AI Global AI Index has returned 113% in the period, compared with 223% for the Global AI Index.

Decisions in passive investing

This comparison is a reminder of the importance of analyzing the composition and methodology of an index in order to understand what the strategy is about, and how it is accomplished.

Working for the right solution to different investors’ needs is at the heart of our business. The AI indices are a good case in point.

Featured indices

1 12-month data as of Feb. 28, 2018.