As concerns about the return of investment in AI gradually emerge, the market's enthusiasm for technology stocks seems to be heading towards differentiation.

According to the research report released by Goldman Sachs' Peter Oppenheimer and Guillaume Jaisson on September 5th, the fundamentals of the technology industry are strong but there is a high concentration risk. It is recommended to seek diversified investments. This not only reduces concentration risk, but also allows investors to enjoy the growth of the technology industry and not miss out on growth opportunities in other industries driven by AI technology.

Technology stocks have not yet become a bubble, and their valuations are in a historically reasonable range.

The report indicates that since the end of the 2008 financial crisis, technology has been the most important driving force for global stock market returns, with profits surpassing other major industries such as media and telecommunications. Considering the strong profit growth of technology stocks, their outperformance of the overall market is relatively reasonable.

The data in the report shows that since 2008, earnings per share (EPS) in the global technology industry have grown by approximately 400%, while the average growth rate in other industries is only about 25%.

Furthermore, the report also states that many historical experiences have shown that the technology behind speculative periods has proven to be transformative, bringing significant secondary innovations, new products and services, and profound social changes in our lives, work, and consumption patterns.

However, the market's excitement over new technologies often evolves into excessive enthusiasm, which can lead to investment bubbles. A recent study found that in 73% of the 51 major technological innovation cases launched between 1825 and 2000, there were clear price bubbles.

However, Goldman Sachs believes that artificial intelligence has not yet become a bubble.

The report points out that the current valuation of the AI industry is far lower than the typical valuations of other recent bubble periods, including the Nifty 50 bubble era in the early 1970s, the Japanese bubble at the end of the 1980s, and the technology bubble of 2000.

The data also shows that the median PE ratio and enterprise value (EV) / sales of the current Mag 7 are only half of the top 7 companies during the dot-com bubble in 2000. Furthermore, the current dominant companies are more profitable than those during the bubble period, and their balance sheets are also stronger.

The report states:

Aggressive new technologies often attract a large amount of capital and competition, but not all events end in a huge bubble. Most end with a downward adjustment in the prices of the entire industry, as the ROI returns to moderate levels.

Ultimately, emerging technology markets often consolidate under a few big winners, and growth opportunities shift to secondary innovations or products and services derived from the technology.

The capital ROI of AI is not that worrisome.

Although the current dominant companies are profitable, it is important to recognize that leading AI companies are no longer light-capital enterprises. They require continuous high capital investment and may even stifle the industry's high ROI advantage over the past 15 years.

This is also a topic that has been widely discussed in the field of AI: high capital expenditure vs. AI technology that has not yet been fully commercialized.

Although capital expenditure remains high, Goldman Sachs' strategy team believes that the return on investment is not so worrisome. During the peak of the technology bubble, capital expenditure and research and development expenses of TMT stocks accounted for over 100% of operational cash flow (CFO). Today, this ratio for TMT stocks is only 72%.

Furthermore, Goldman Sachs believes that the significant increase in capital expenditure may actually generate strong returns.

For example, from 2013 to 2016, Microsoft made massive capital expenditures to build Azure. Azure's gross margin was initially negative, but later became a huge profit.

Compared to Azure, Microsoft's generative AI has a faster profit growth rate (annualized $0.5-6 billion), while it took Azure about 7 years to reach a comparable level.

In other words, although the capital intensity of top AI companies has increased significantly, their revenue has quickly reached a comparable level to Azure's.

There is a high concentration of risk, so it is recommended to focus on the ETC theme.

The data shows that although the valuations of top technology companies are not as high as those of previous bubble periods, their market share is the highest it has been in decades, accounting for 27% of the total market cap of the S&P 500, indicating an unprecedented level of concentration.

The report compiled the average total return that can be obtained by buying and holding the top ten stocks within 1-10 years since 1980. It was found that although the absolute returns of dominant companies are still good, these strong returns gradually diminish over time and they often remain stable "compound companies".

More importantly, if investors buy and hold advantaged companies while other fast-growing companies emerge and outperform the market, the returns of advantaged companies usually turn negative.

Considering the above factors, Goldman Sachs believes:

1) Dominant companies are unlikely to be the fastest-growing companies in the next ten years.

2) The specific stock risks in the current index are very high, and diversification can increase returns.

Goldman Sachs found that in addition to the technology industry, other industries also have companies with high profit margins and investment returns, as well as strong balance sheets. The report named them "Ex Tech Mixers (ETC)".

Specifically, the global ETC list needs to meet the following conditions:

Market capitalization exceeding $10 billion, high profit margins (EBITDA > 14%, EBIT > 12%, net profit > 10%), high profitability (ROE > 10%), strong balance sheet (net debt/equity < 75%, ND/EBITDA < 2x), low volatility (vol < 50), and sustained profit growth (sales > 4%, profit > 8%), with continuous profit growth over the past decade.

The report statistics found that ETC's performance in the past year was better than the global market, and it stayed in sync with the performance of the "Mag 7". The valuation is consistent with the average level since 2016, and the trading price relative to the world stock market is at its lowest level since 2018.

Specifically, the sectors mainly covered by ETC include: medical care and biotechnology, banks and financial services, consumer goods and services, siasun robot&automation and cybersecurity, retro consumer themes, traditional economy and infrastructure.