Artificial intelligence stocks were stock market darlings in 2024. Investors clamored to add AI stocks to portfolios alongside a tidal wave of investment to train and operate AI chatbots and agentic AI software solutions.
AI stocks like Nvidia and Palantir surged 171% and 340%, helping lift the S&P 500 index by 24% last year.
So far, those heady returns haven’t continued into 2025, especially recently. The S&P 500 is down about 2%, and many technology stocks have tumbled much more. Nvidia is down 16% year to date, while Palantir has retreated 37% since mid-February.
The launch of the Chinese AI chatbot DeepSeek on January 29th raised questions about the amount of AI spending, the competitive position of U.S. AI companies, and, ultimately, the amount of continued investor demand for AI-related names in the stock market.
Related: Rare event could send Amazon stock surging after big drop
Investors are likely scratching their heads, wondering what’s next.
Recently, Goldman Sachs weighed in on the matter, laying out the AI stocks it thinks offer opportunity. Overall, the investment firm believes investors should rotate away from what it calls “Phase 1” stocks like Nvidia and “Phase 2” stocks like AI infrastructure stocks. Instead, it likes stocks likely to generate AI-enabled revenues, or “Phase 3” stocks.
AI stocks have tumbled, prompting Goldman Sachs to refresh its AI stocks wish list.
Michael M. Santiago/Getty Images
How have Goldman Sachs models done recently?
Since the S&P 500 market peaked on February 19th, the Goldman Sachs AI Phase 2 basket has declined 14%, the Goldman Sachs AI Phase 3 basket is down 12%, and the Goldman Sachs AI Phase 4 basket has fallen just 7%. This compares with a 4% decline for the equal-weighted S&P 500.
What May Turn Things Around?
What will it take to see some improvement in these names? The team at Goldman Sachs believes that a combination of washed-out sentiment and better economic data will likely be the catalysts for the group’s better performance.
Sentiment is certainly negative right now. However, it is not as washed out as Goldman Sachs typically sees before things start to turn around. Ultimately, longer-term gains for the group will likely be driven by a combination of “technological progress and earnings growth.”
When large hyperscaler companies like Amazon’s AWS recently reported results, they again raised their capital expenditure outlooks for 2025. However, capex trends will likely start slowing year over year, especially if the U.S. economy weakens or capacity begins to outpace demand.
In addition, investors are increasingly asking whether the large amounts these companies are spending on capex are really going to generate positive returns on investment down the road.
Where are we in terms of AI adoption?
One of the most helpful pieces of information comes from the U.S. Census Bureau.
A recent survey indicates that 5% of U.S. companies are using AI to produce goods and services. However, a recent report from the Federal Reserve Bank indicates that the number is closer to 20% (with some adjustments, such as employment-weighting the survey). This latter estimate is similar to a variety of other, smaller surveys published on AI adoption.
What criteria does Goldman Sachs use for its updated ‘Phase 3’ AI stocks?
Goldman Sachs has updated its Phase 3 list of names utilizing the following criteria:
1) Stocks that have traded with a statistically significant beta to the firm’s AI Phase 3 basket over the past year after controlling for market beta and interest rates (note: beta represents the market volatility of a security versus an underlying index)
2) Stocks that have mentioned revenue-enhancing AI use cases in their most recent conference call.
There are 40 names on the list, and 22 are new stocks. Of the 40 names on the list, 27 are software companies, and the ones currently expected to generate the strongest revenue growth over the next two years include:
Palantir Technologies ( (PLTR) ) – Palantir is a data analytics company. The firm serves commercial and government clients via its Foundry and Gotham platforms. Earnings per share are forecast to increase 36% this year and 26% next year, and the company’s forward 12-month P/E currently stands at 137x. There is currently 40% upside to the average analyst target price for the stock.
Cloudflare Inc. ( (NET) ) – Cloudflare is a security and web performance solutions software company with a distributed, serverless content delivery network, or CDN. Its edge computing platform allows users to deploy and execute code without maintaining servers. Earnings per share are forecast to increase 6% this year and 30% next year, and the company’s forward 12-month P/E currently stands at 146x. There is currently 13% upside to the average analyst target price for the stock.
Sentinel One Inc. ( (S) ) – SentinelOne is a cloud-based cybersecurity company focused on endpoint protection. SentinelOne’s Singularity platform helps enterprises detect and respond to IT security threats. Earnings per share are not available for this year but are forecast to grow 853% from $.02 to 20 cents next year. The company’s forward 12-month P/E currently stands at 93x. There is currently 50% upside to the average analyst target price for the stock.
GitLab Inc. ( (GTLB) ) – GitLab offers DevOps point solutions and DevOps platforms. GitLab is available via self-managed and software-as-a-service (SaaS) models. Earnings per share are forecast to decline 1% this and increase 32% next year and the company’s forward 12 month P/E currently stands at 70x. There is currently 40% upside to the average analyst target price for the stock.
Related: Veteran fund manager unveils eye-popping S&P 500 forecast