Interest in generative artificial intelligence is intensifying, as measured by internet search volumes, news stories, and more discussion of the topic than ever before on fourth-quarter earnings calls. The economic potential of the new technology has helped lift the stock market to new highs, led by Nvidia, the maker of specialized chips that are crucial to run generative AI models.
Now a key question for investors is: What happens next as the AI rally broadens to involve more companies? According to Goldman Sachs Research, if Nvidia represents the first phase of the AI trade, Phase 2 will be about other companies that are helping to build AI-related infrastructure. Phase 3 deals with companies incorporating AI into their products to boost revenue, while Phase 4 is about the AI-related productivity gains that should be possible across many businesses.
Are AI stocks in a bubble?
So far, investor optimism isn’t running as high as it was at prior peaks in 2000 and 2021, Goldman Sachs Research strategist Ryan Hammond writes in the team’s report. The implied level of long-term earnings growth that investors expect has climbed to 11% a year. That’s above the long-run average of 9% but still below the 16% percent growth that was expected at the height of the technology bubble in 2000 or the 13% growth implied by stock prices at the peak of the post-Covid rally in 2021. The analysis is based on the relationship between return-on-equity and price-to-book, to quantify the long-term growth in earnings implied by how the market is priced today.
As another way to measure optimism, analyst growth estimates are optimistic for the 10 largest technology companies but, again, are below what was seen during the technology bubble. The median large tech company is expected to deliver 15% earnings per share growth in three years, according to analysts’ consensus estimates. That compares with a median of 11% for the S&P 500 as a whole, but it’s well below the 24% earnings growth analysts foresaw in March 2000.
What’s more, current valuations of large tech stocks are not as stretched as they were in prior periods. The 10 largest are trading at 28 times earnings, which “pales in comparison to the peak of the tech bubble,” Hammond writes. Valuations for the 10 biggest tech companies reached 52 times earnings in 2000 and touched 43 times in late 2021.
What’s next for AI investment in the stock market?
Phase 1: Nvidia
Nvidia shares have returned more than 500% since the start of 2023. Remarkably, though, those gains have been entirely driven by earnings growth: The company’s price-to-earnings ratio is barely higher than it was at the start of last year, according to Goldman Sachs Research.
While Nvidia has dominated the initial phase of the AI trade, the companies in Goldman Sachs Research’s three subsequent phases have also outperformed relative to their industry groups over the past six months.
Phase 2: AI infrastructure
Companies beyond Nvidia that stand to benefit from the buildout of AI-related infrastructure include semiconductor designers and manufacturers, cloud providers, computer and network equipment makers, data center real estate investment trusts, utilities, and security software providers.
To date, the performance of stocks in this group has varied widely, Hammond writes. Companies that design chips and have intellectual property in this area, along with security companies, have had significant gains. On the other hand, utilities that might be expected to benefit from increased electricity demand from data centers have barely budged.
Phase 3: AI-enabled revenue
Companies that can incorporate generative AI advances into their product offerings make up the next phase of the AI trade. Software and IT services stocks may be best positioned, according to Goldman Sachs Research, and many of them have begun to describe for investors how their tools will enable other companies to use the new AI technology.
The research team identified stocks that fall into this group by screening for companies that made relevant mentions of AI product offerings during fourth-quarter earnings calls. They also screened for shares that trade with a statistically significant beta to Nvidia stock, suggesting they are reacting similarly to news and market developments.
“The outperformance of these stocks, while driven by many factors other than exclusively AI, suggests investors have begun to trade Phase 3 already,” Hammond writes.
Phase 4: AI-productivity gains
Eventually, emerging AI technology can be expected to benefit companies across a range of industries that can use it to boost productivity.
Software and services companies and commercial and professional services firms appear to have the biggest potential for earnings gains from AI, because they have a combination of relatively high labor costs overall and a high share of their labor bill that may be exposed to AI automation.
The report notes that 30% of all mentions of AI-related productivity gains on fourth-quarter earnings calls came from companies in these industries. Such mentions likely signal both strong opportunities for AI to boost productivity and the ready willingness by the management team to pursue and embrace the potential.
Where are we now?
So far, few companies come close to reflecting AI optimism the way Nvidia has. But stocks in Phases 2 and 3 have shown more signs of investor optimism than companies in Phase 4, the report suggests. “Many companies in these two phases are necessary for every other company to use the technology to improve productivity,” the report notes.
This article is being provided for educational purposes only. The information contained in this article does not constitute a recommendation from any Goldman Sachs entity to the recipient, and Goldman Sachs is not providing any financial, economic, legal, investment, accounting, or tax advice through this article or to its recipient. Neither Goldman Sachs nor any of its affiliates makes any representation or warranty, express or implied, as to the accuracy or completeness of the statements or any information contained in this article and any liability therefore (including in respect of direct, indirect, or consequential loss or damage) is expressly disclaimed.