David Trainer, founder and CEO of investment research firm New Constructs, told Fortune earlier this summer that he fears investors are getting ahead of themselves by betting on richly valued A.I.-linked equities. “We’re seeing just one fad after another. It’s FOMO, and more and more stocks are moving to ridiculous heights…Investors just have to be really careful,” he warned. 

But Goldman Sachs analysts, led by vice president of U.S. equity strategy Ryan Hammond, said they remain bullish on A.I. in a Monday research note.

“Our economists believe A.I. should make workers more productive, increasing corporate revenues. Alternatively, A.I. adoption could allow some companies to generate the same amount of revenues but with lower labor costs that would increase margins,” Hammond and his team wrote.

The analysts argued that although the timing of broad-scale A.I. adoption is still “highly uncertain,” the technology should have “a meaningful macro impact some time between 2025 and 2030” and begin boosting corporate earnings even before then.

For stock market investors, that means opportunity, so Hammond and his team highlighted key companies that are set to benefit from A.I. in the short- and long-term. Here’s how they broke it down.

The short-term winners: Chipmakers and data centers

First, in the short term, Goldman named three categories of A.I. winners: The so-called enablers, hyperscalers, and empowered users.

“Enablers” include key producers of the underlying hardware that allows A.I. to operate, such as semiconductors and related equipment. These are the “picks and shovels” of the A.I. “gold rush.” Goldman highlighted the semiconductor makers Nvidia and Marvell Technology, as well as Credo Technology Group, which provides electrical cables and other hardware for data centers, as their top “enabler” picks for investors.

Then there are the “hyperscalers.” These are big tech firms—including Microsoft, Alphabet, and Amazon—whose cloud-computing divisions should benefit from the large-scale commercialization of A.I. Generative A.I. chatbots and related technologies rely heavily on cloud computing resources and infrastructure to train and operate.

Finally, there are the “empowered users”—technology companies that are already using A.I. to expand their product and service offerings. The list includes the social media giant Meta Platforms, which is using A.I. to improve targeted advertising and has multiple generative A.I. models of its own, and software provider Adobe, which has incorporated A.I. into Photoshop. The financial software firm Intuit as well as client relationship management (CRM) software giants Salesforce and ServiceNow also made the list.

Long-term: More output, smaller workforce

Over the long term, Goldman expects A.I. will not only boost worker productivity, but also allow some firms to slash their headcounts. These effects combined should reduce labor costs and increase revenues, lifting earnings. 

Hammond and his team didn’t put out a timeline for when they expect widespread adoption of A.I. to lift the corporate bottom line, but they did make a pretty optimistic forecast for when it does.

“Our framework implies earnings for the median Russell 1000 stock could be 19% greater than the baseline via widespread A.I. adoption and increased labor productivity,” they wrote, referencing the index that tracks the top 1,000 companies by market cap in the U.S.

Big companies seem to have bought into those predictions. Despite recent suggestions the A.I.-linked stock surge could be a bubble, corporations are betting big on the technology. From publicly traded telecom giants like AT&T to VC funds like Bessemer Venture Partners, billions have been spent this year alone on A.I. initiatives and investments. And even though the public hype over ChatGPT has calmed, corporate executives still have A.I. on their mind. 

Just look at the steady rise in references to the technology on corporate earnings calls. Even though there’s still a week remaining in August, mentions of the word “A.I.” are already at a monthly high.

For investors looking to get into long-term A.I. winners, Goldman recommended focusing on two key characteristics. The firms that will benefit most from A.I. have high labor costs relative to their revenue and can use A.I. to reduce a large portion of those costs. 

Typically, these are going to be companies in industries with a lot of white-collar workers, including the software, consumer staples, and professional services sectors, according to Goldman analysts.

Using this framework, Goldman produced a group of 50 companies they believe will be “long-term A.I. beneficiaries.” Here’s the full list: