There's little doubt as to the stock market's biggest winners as it relates to artificial intelligence (AI). Nvidia's (NVDA 4.35%) processors are the centerpiece of most AI data centers, while Palantir Technologies' artificial intelligence software is making the most out Nvidia's hardware. Both stocks have performed incredibly well over the course of the past couple of years.

As is the case with any young industry, though, smaller companies are finally figuring out how to capitalize on AI's growth. Although outfits like Palantir and Nvidia are at no immediate risk of being dethroned in terms of size or reach, don't be surprised to see some new names taking the lead in terms of performance as we begin the next chapter of the AI era.

Here are three such names to put on your radar, if not yet in your portfolio.

An AI robot pressing a virtually displayed button.

Image source: Getty Images.

1. CoreWeave

It's an underappreciated nuance of the business. But most companies utilizing artificial intelligence technologies don't actually own their own AI data centers. It's often cheaper and easier to simply rent cloud-based access time to these platforms. That's what CoreWeave (CRWV 9.63%) offers.

Granted, cloud-accessible AI platforms aren't exactly uncommon anymore. As was noted, in fact, it's the norm within the business. CoreWeave's tech is still special though in that the company's infrastructure is purpose-built from the ground up to offer access to the industry's newest and most powerful processors, like Nvidia's GB200 Blackwell Superchip, which trains AI models 4 times faster than previous processing platforms, and performs real-time inference calculations up to 30 times faster than most platforms currently in use can. Indeed, to call CoreWeave's solutions "fifth-generation" arguably still understates the performance leap it offers its customers -- customers that include OpenAI, Microsoft, and Meta Platforms, by the way.

So why haven't more investors heard of it? Although it's not a new enterprise, it is a relatively new publicly traded stock. The company's initial public offering (IPO) was in March of this year, when investors were far more concerned about the tariff talk that was upending the market.

Regardless, although it's currently unprofitable, its first-quarter revenue soared 420% to nearly $1 billion. Analysts expect triple-digit top-line growth this year and next, with a swing to profitability likely by 2027. CoreWeave is obviously coming into its own now that technology giants recognize how much more they can accomplish with next-generation AI solutions.

2. Astera Labs

With nothing more than a passing glance, Astera Labs (ALAB 2.62%) looks a lot like CoreWeave, seemingly making it unnecessary to own both.

The two companies are actually quite different, though. CoreWeave's business model offers a turnkey solution to institutions that want access to next-generation AI computing platforms. Astera simply makes and sells much of the hardware you'll find within a data center's racks of processors connected to circuit boards, each of which is connected to another. And like CoreWeave -- and unlike much of the hardware used in artificial intelligence's early days -- much of Astera Labs' technology is purpose-built specifically for AI.

Case in point: Astera's Aries line of PCIe/CXL connectivity tech supports data transmission between AI processors at speeds of up to 64 gigatransfers (64 billion transfers of digital data) per second. And it does so using Ethernet cables that are much thinner than the ones you might normally use in a home or office setting. That may not mean much to the average person, but this will: This option allows for more computing power with less power consumption and a smaller physical footprint. This ultimately translates into lower operating costs. The company also offers PCIe switches, digital signal processors, and memory controllers that similarly offer better functionality in a smaller package.

But are these technological solutions the industry actually needs enough to pay a premium for? This might answer the question: Astera's revenue is expected to grow to the tune of 80% this year, pushing the company out of the red and into the black. And that's just the beginning of the growth streak analysts expect of this fast-growing up-and-comer.

3. Penguin Solutions

Finally, add Penguin Solutions (PENG 1.22%) to your list of overlooked AI stocks with room to run.

You've probably never heard of it, but not because it's new -- the company's been publicly traded since 2017, when it was still called SMART Global Holdings. The name change in 2024 probably isn't the reason you aren't familiar with it either. Rather, the chief reason you've probably never come across this outfit is its sheer size. With a market cap of only $1.3 billion, Penguin is not only the smallest of the three AI names in focus here, but it's just plain small.

Don't let its diminutive size dissuade you, though. This outfit packs a powerful growth punch. The analyst community is looking for sales growth of 17% this year, mostly driven by growing demand for -- surprise, surprise -- its AI data center infrastructure solutions.

Once again, with just a quick look, Penguin looks a lot like Astera and CoreWeave. And there's some overlap to be sure. Penguin Solutions is different enough, though. It offers some of the specialized solutions somewhere in between Astera Labs' tech and CoreWeave's service, like fault-tolerant computing memory modules, data center design and deployment services, and software specifically meant to manage the computing clusters that most modern AI data centers now utilize.

The kicker: Unlike so many of its fellow AI outfits -- and especially the smaller ones -- Penguin Solutions is consistently profitable. Moreover, the stock's only trading at about 15 times this year's projected per-share profits of $1.62, which should grow 17% to $1.89 next year. Not bad for such a little outfit.

Just remember that it is a small-cap stock capable of dishing out above-average volatility. You'll want to handle it accordingly.