The AI Trade Is Splitting in 2: Here Are the Winners and Losers

Artificial intelligence has been the driving force behind the current bull market, but some cracks are starting to appear in the megatrend. The entire industry has been under pressure over the last month, and there's a growing divide in outlooks across various segments of the AI market. That divide could grow wider over the next few quarters, creating clear winners and losers.

The days when any company related to artificial intelligence could find success in the stock market may be coming to an end. Here's what investors need to know.

One of the biggest bottlenecks in artificial intelligence is getting tighter

Perhaps the biggest bottleneck for most companies building new data centers is access to capital.

Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL), Amazon, and Microsoft have funded much of their build-out using free cash flow from their massive businesses outside of cloud computing. All three, however, have recently been pushed to raise more outside capital to fund their build-outs, with Alphabet and Amazon producing negative free cash flow in recent quarters.

The neocloud companies that stand up data centers more quickly and offer simple infrastructure as a service are almost fully reliant on debt. CoreWeave (NASDAQ: CRWV) may be the worst of the bunch. It held $51.4 billion in debt and lease liabilities on its balance sheet, as of the end of the second quarter, with just $6.4 billion in cash.

Rising interest rates and declining confidence in the returns for AI spending have led to a rising cost of capital for everyone in the industry. S&P Global recently revised Meta Platforms' outlook to "negative" from "stable" due to the uncertainty around its AI spending. That's pushed the spread it pays above the risk-free rate on new debt significantly higher.

But the rise in Meta's debt pricing isn't nearly as much as the neoclouds. Oracle, for example, has seen its spread widen from less than 200 basis points to 275 basis points over the last two months. Yields on CoreWeave's debt have topped 15% recently, and it, along with other neocloud companies like Nebius, have been effectively shut out of the bond market, according to a report from the Wall Street Journal.

While the hyperscalers might have to pay a little bit more for access to capital, at least they have access to one of the most important resources for their continued growth.

The election results could add to the AI divide

Voters elected Democrat governors and representative majorities in several key states, which could result in policies that grow the divide between hyperscalers and neoclouds. Democrats had a surprising victory in Texas, which could have governor-elect Gina Hinojosa extend the temporary moratorium on data centers imposed by outgoing Republican governor Greg Abbott.

Texas is the state with the most planned data center construction. Democrats also now control the executive branch in the three states that follow Texas: Ohio, Georgia, and Arizona. UBS analysts estimate about 30% of expected 2028 capacity additions are at risk of delays following the election.

That increases the value of existing data centers and projects already cleared and funded, such as those owned by the hyperscalers. Meanwhile, the neoclouds, which typically build faster, could see significant delays in expanding capacity while their debt loads continue to weigh on earnings. That will grow the divide between the two groups.

It's worth noting that neoclouds aren't the only companies that stand to lose. Chipmakers like Nvidia (NASDAQ: NVDA) will also see a negative impact if neoclouds aren't able to build as quickly as planned. That's especially true considering neoclouds typically stick with Nvidia chips for most of their servers. Meanwhile, the hyperscalers use custom silicon chips for a growing share of their servers due to the capital efficiency and performance they provide for their own AI development and their customers. That's been reflected in the sell-off in semiconductor stocks over the past month or so.

Investors looking to maintain a position in AI stocks should consider reducing exposure to neocloud companies and GPU makers while maintaining exposure to the hyperscalers. In fact, the recent sell-off across tech stocks may be presenting buying opportunities among hyperscalers right now.

Adam Levy has positions in Alphabet, Amazon, Meta Platforms, Microsoft, and S&P Global. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, Nvidia, Oracle, and S&P Global. The Motley Fool has a disclosure policy.