This summer, most stocks experienced an abrupt correction, including many winners in the artificial intelligence (AI) sector that had outperformed in the first half of the year. Economic fears have driven almost all stocks down since mid-July.

But even if there is an economic downturn, investment in AI should continue. And if the economy weakens, that should lead to lower interest rates, which in turn should keep valuations of technology growth stocks high.

That means stocks that could benefit from artificial intelligence, like these three leaders, could be great buys or complementary stocks during a market-wide decline.

Broadcom

Broadcom (AVGO -0.25%) CEO Hock Tan has masterfully implemented a growth strategy through acquisitions, focusing on leading companies in certain niche technologies. And two specific technologies in Broadcom’s chip portfolio have taken off with AI.

These include Broadcom’s market-dominating Tomahawk and Jericho switching and routing chipsets, which are experiencing rapid growth due to the high data transport requirements of AI.

The second AI business is custom ASICs (application-specific integrated circuits), which are used by both alphabet (GOOG 0.96%) (GOOGL 1.03%) And Meta-platforms (META -1.84%) to make their own custom AI accelerators. Broadcom recently announced a third major ASIC customer for AI accelerators, which some analysts believe is TikTok parent company Bytedance.

These two AI businesses have exploded, growing from just over $4.2 billion last year to what is expected to be more than $11 billion this fiscal year, which ends in October. Broadcom also has other highly profitable businesses, such as providing Wi-Fi and Bluetooth chips for the iPhone.

In addition to chips, Broadcom’s recent acquisition of software giant VMware is also proving to be a resounding success. By integrating the virtualization software giant into its business, Broadcom was able to cut costs and increase revenue, significantly boosting VMware’s profits.

With the company’s business now almost evenly split between software and hardware – a rarity among major technology stocks – Tan can now look to either hardware or software companies for his next acquisition. That’s an advantageous position in the rapidly evolving AI space.

Down about 15% from its highs and with a price-to-earnings ratio of just 25, Broadcom is a solid AI pick to buy on dips.

ASML Holding

Another AI winner is ASML Holding (ASML -1.03%)which is fortunate enough to have a monopoly on extreme ultraviolet lithography (EUV) technology. EUV is required for the manufacture of all modern semiconductors and is currently used in the production of advanced DRAM memory – which is also critical in AI systems.

Because of its competitive advantage, ASML never really trades “cheap,” but with the stock down 23% from its recent highs, this could be a good time to buy shares.

While ASML’s second-quarter results beat expectations, they may have disappointed some who had hoped for shorter-term growth. But ASML’s revenues can fluctuate. Management has always indicated that 2024 will be a year of growth pause before a likely big year in 2025.

Last year, the rest of the chip sector, with the exception of AI, was still in a post-COVID crisis. But this year, AI continues to grow rapidly and now accounts for a larger part of the industry. At the same time, the AI ​​revolution is expected to increase the chip share in more devices such as AI-enabled PCs and smartphones. This should boost industry-wide growth through 2025.

Recently, there have been signs of a sharp increase in demand for chips. Semiconductor manufacturing in Taiwan (TSM 0.33%)the largest foundry for cutting-edge chips and ASML’s largest customer, has just reported an astonishing 44.7% increase in sales in July compared to last year and a 30.5% increase in sales year-to-date.

With such rapid growth, TSMC will likely increase spending again soon, which should mean higher revenues for ASML and thus enable the company to make a strong recovery from this decline.

Image source: Getty Images.

Super-microcomputer

Regardless of which GPU vendor wins out for advanced AI applications, future AI GPUs will likely need to be liquid-cooled in a server rack. But that’s a tricky proposition because direct liquid cooling (DLC) technology has been an expensive and cumbersome proposition to date, which is why 99% of all data centers will be air-cooled by 2024.

However, the enormous power requirements of AI are now forcing data center operators to introduce DLC. And the server manufacturer that has led this trend is Super-microcomputer (SMCI 0.34%).

Super Micro missed earnings estimates last quarter, and the stock plummeted accordingly. It is now 56% below its March all-time high, although the stock is still up 100% from the start of 2024.

But if you look behind the scenes, the reasons for the profit shortfall are not so worrying. Super Micro already has a working DLC ​​server product, and AI customers are clamoring for it. Demand was so high that management had to cover the cost of express shipping liquid cooling components, which squeezed margins. The supply shortages also squeezed revenue by $800 million from the quarter that ended in late June to the quarter that ended in late September.

Despite the shortage, CEO Charles Liang estimated that Super Micro shipped 1,000 liquid-cooled racks in June and July, representing about 15% of all data center deployments and “at least” 70 to 80% of all DLC servers worldwide, giving the company dominance in this emerging market.

All of these seem like valid reasons for the quarter’s miss, especially given Super Micro’s recent track record of exceeding all expectations. Meanwhile, management is forecasting a gradual improvement in the company’s gross margins, which will return to the target range of 14% to 17% by year-end.

Super Micro’s management also gave a revenue forecast of $26 billion to $30 billion for next year, compared to the nearly $15 billion the company generated in the 12 months through June. And Super Micro has a history of conservative forecasting. A year ago, for example, Super Micro forecast revenue of $9.5 billion to $10.5 billion for the 2024 fiscal year ending in June. The company ended up generating 50% more than that figure.

This suggests that Super Micro could come in well above expectations in terms of revenue, which could help earnings beat expectations in 2025 despite some margin compression. Given that the stock trades at only about 16 times conservative earnings estimates for next year, it’s a buy on this dip.