The artificial intelligence (AI) market has grown like a weed over the past decade. That rapid expansion -- which was fueled by more sophisticated cloud computing services, large language models, and generative AI applications -- lit a blazing fire under some high-growth tech stocks.
The most obvious winners are Nvidia, the world's top producer of discrete graphics processing units (GPUs) for processing AI tasks; and Microsoft, which acquired a big stake in OpenAI and integrated the start-up's AI tools into its own services. But there are still plenty of other under-the-radar AI growth plays that might have more upside potential. Let's look at two of them: Credo Technology (CRDO +0.10%) and Arm Holdings (ARM 0.02%).
Image source: Getty Images.
Credo Technology
Credo, which went public in 2022, sells a wide range of high-speed connectivity solutions for the data center, cloud, and AI markets. Its core products include data transfer chips, digital signal processors, line card components, and active electric cables for data centers. From fiscal 2022 to fiscal 2025 (which ended this May), Credo's revenue grew at a compound annual growth rate (CAGR) of 60%. It also turned profitable for the first time in fiscal 2025.
NASDAQ: CRDO
Key Data Points
Credo attributed that growth to the rapid expansion of the cloud and AI markets, which drove its hyperscale customers to aggressively upgrade their data center infrastructure. Its biggest customer -- widely believed to be Microsoft -- accounted for 39% of its revenue in fiscal 2024. That customer concentration isn't ideal, but it also isn't surprising considering how much Microsoft is prioritizing the expansion of its cloud and AI ecosystems. Its other major customers include Amazon and Tesla.
From fiscal 2025 to fiscal 2027, analysts expect Credo's revenue to rise at a CAGR of 47% as its earnings per share (EPS) increase at a CAGR of 113%. That rapid growth could be driven by the continued expansion of the AI market, and a shift toward higher-speed ethernet connections that will spur demand for its new optical modules. There's also rising demand for its "chiplet" designs, which are more modular, customizable, and scalable than monolithic system on chips (SoCs), which merge together multiple chips on a single die. It isn't cheap at 64 times this year's earnings, but it could have plenty of room to run over the next few decades.
Arm Holdings
Arm is a U.K. chip designer which was acquired by Japan's SoftBank in 2016 and spun off again in a second IPO in 2023. It develops power-efficient CPUs that consume less power than the x86 CPUs produced by Intel and AMD.
That makes Arm's chips well-suited for smartphones, tablets, Internet of Things (IoT) gadgets, connected vehicles, and even some notebook computers and servers. Its chip designs are now installed in approximately 99% of the world's smartphones, and most of its growth over the past few years has been fueled by its AI-optimized Armv9 designs.
NASDAQ: ARM
Key Data Points
Arm's revenue rose 24% in fiscal 2025 (which ended this March), and analysts expect that figure to grow at a CAGR of 21% over the next three years. Its EPS, which surged 159% in fiscal 2025, is expected to grow at a CAGR of 41% through fiscal 2028.
Arm originally only generated its revenue by licensing its designs to chipmakers like Qualcomm, MediaTek, and Apple instead of producing its own chips. But earlier this year, it announced that it would start developing its own first-party chips and outsource its production to Taiwan Semiconductor Manufacturing. That surprising move would boost Arm's operating expenses and turn it into a direct competitor for some of its top clients. But it could also undercut other Arm-based chips because it doesn't need to pay any royalties or licensing fees for its own designs. Its first-party brand recognition could also make it a more appealing option than third-party Arm-based chips for OEMs.
Arm's stock certainly isn't cheap at 113 times this year's earnings, but it could be a great long-term play on the market's growing demand for more power-efficient AI chips.
