It is easy to think of artificial intelligence investing as a mega-cap story. After all, the AI boom has been led by some of the largest companies in the market, from chipmakers like Nvidia Corp. (ticker: NVDA) to cloud platforms such as Microsoft Corp. (MSFT) and Alphabet Inc. (GOOGL, GOOG). But investors looking for the best AI stocks of the future may want to look one layer deeper.
“The AI trade is moving from the spenders of cash to the receivers of cash,” says Tejas Dessai, director of thematic research at Global X. In the first phase, he says, investors focused on the mega-cap platforms building AI capacity. “Right now, the most urgent opportunities are in the physical constraints that determine how quickly AI capacity can come online.”
[Sign up for stock news with our Invested newsletter.]
In other words, the companies writing the biggest checks for AI may not be the only ones that benefit. Smaller companies can also gain if they provide the equipment, infrastructure or specialized services needed to advance AI.
Still, investors should tread carefully. Many emerging AI companies are speculative, and small-cap stocks can be volatile. Dessai says the more promising names should address constraints in areas such as memory, power, grid infrastructure, cooling or automation, while also showing customer demand through backlog growth, contracts or long-term infrastructure commitments.
There are two primary ways to tap into these areas: For a diversified approach that doesn’t require doing your own company research, thematic exchange-traded funds, or ETFs, are a great approach. For example, the Global X Artificial Intelligence & Technology ETF (AIQ) targets companies involved in AI development and implementation across hardware, software and supporting technologies. Or the Global X U.S. Electrification ETF (ZAP) focuses on companies tied to U.S economy modernization and electrification, something AI has a ravenous appetite for.
Another approach is to buy individual stocks. This gives you the greatest control over your portfolio but comes with the most risk and research requirements. If you’re willing to take this on, the following six small-cap AI stocks and emerging AI companies offer different ways to play the next stage of the AI boom:
| Stock | Market Cap | Why It Stands Out |
| SkyWater Technology Inc. (SKYT) | $1.8 billion | U.S. chipmaking capacity |
| Arteris Inc. (AIP) | $1.7 billion | AI chip design tools |
| Preformed Line Products Co. (PLPC) | $1.8 billion | Grid infrastructure demand |
| POET Technologies Inc. (POET) | $2.3 billion | Solves speed bottlenecks using light |
| QuickLogic Corp. (QUIK) | $383.7 million | Programmable chip technology |
| Innodata Inc. (INOD) | $3.2 billion | AI training |
SkyWater Technology Inc. (SKYT)
SkyWater Technology may not be the flashiest AI stock, but this is part of what makes it so appealing for savvy investors. The U.S.-based semiconductor manufacturer is on the frontline of one of several AI bottlenecks: First and foremost, it feeds AI’s insatiable appetite for powerful chips. But AI also needs reliable supply chains, domestic production and specialized manufacturing know-how. SkyWater provides all of that.
Its latest earnings release shows just how valuable these services are. The company reported record revenues and 29% year-over-year growth for fiscal year 2025. This also drove record gross profit, net income to shareholders and adjusted earnings before interest, taxes, depreciation and amortization, or EBITDA. Part of this revenue increase was due to completing eight engagements with quantum computing companies and its midyear acquisition of Fab 25, which greatly increased SkyWater’s operational capacity.
There is also a timely twist to this investment case: Quantum computing firm IonQ Inc. (IONQ) is set to acquire SkyWater this year in a roughly $1.8 billion cash-and-stock deal. This broadens SkyWater’s investment thesis from a pure AI software story into an emerging infrastructure play behind next-generation computing.
Arteris Inc. (AIP)
Arteris tackles one of AI’s less visible problems: how to move data quickly and efficiently inside complex chips. The company provides semiconductor system intellectual property, including network-on-chip technology, that helps chip designers improve chip performance and productivity. And as the energy grid will tell you, AI chips are only useful if they can move data without wasting time or power.
Arteris also has the kind of customer validation investors like to see. The company reported record annual contract value plus royalties of $92.8 million in the first quarter of 2026. This was up 39% year over year. Revenue also reached new highs of $22.9 million for the quarter. To top things off, Arteris and processor IP company MIPS recently announced a partnership to accelerate the buildout of AI computing platforms.
Preformed Line Products Co. (PLPC)
Preformed Line Products may not look like an obvious AI stock at first glance, but it fits neatly into one of the more important parts of the AI buildout: keeping the lights on. The company makes products used in energy, communications and grid infrastructure, giving it exposure to the utility upgrades needed as data centers, factories and other power-hungry facilities expand.
That lines up with where Dessai sees opportunity. Within the broader AI infrastructure theme, he says, “U.S. electrification and U.S. infrastructure development stand out as two of the more compelling themes over the next 12 months, as AI demand reinforces already-strong investment cycles.”
Preformed Line Products also has evidence of demand. First-quarter net sales rose 19% year over year, helped by higher U.S. demand for energy and communications products. The company’s gross profit margin was 31.3%, up 1.5% sequentially from Q4 2025.
POET Technologies Inc. (POET)
As AI workloads grow, data centers need faster ways to move information between chips, servers and racks. POET aims to resolve this bottleneck by using optical technology to move data with light.
The company develops photonic integration technology, including its proprietary Optical Interposer platform, for AI networks. It’s not bringing in consistent profit yet, with a 2026 first-quarter net loss of $12.3 million, largely due to high research and development costs. But brighter times may be ahead: POET recently announced a collaboration with Liteon Technology, an optoelectronic semiconductor component and high-power grid optical system provider. Together they hope to develop the next-generation optical communication modules using POET’s patented technology and platform.
POET also plans to move its headquarters and domicile to the U.S. so it won’t be considered a foreign corporation anymore or a passive foreign investment company (PFIC) in the future.
QuickLogic Corp. (QUIK)
QuickLogic develops technology that helps customers create chips that can be programmed for specialized uses after they’re made. Customers range from defense to AI applications. That matters because not every AI workload will run in a massive data center. Some will need smaller, more efficient chips closer to where data is created.
The stock is still a speculative play, but it shows signs of momentum. The company was recently awarded a seven-figure contract for programmable chips to be used by defense and commercial customers. This should generate revenue for the company from second quarter 2026 through first quarter 2027. Revenue is already on the upswing as of the first quarter of 2026, when the company reported a 16.8% increase in revenue compared to the prior-year period.
Innodata Inc. (INOD)
With a market capitalization above $3 billion, INOD doesn’t fit the small-cap criteria of being between $250 million and $2 billion. However, it still belongs in the emerging AI company camp.
Innodata helps AI builders with the less glamorous but essential work behind artificial intelligence: data engineering, training data, model evaluation and human expertise. This makes the company a different kind of AI pick than a chip manufacturer or data center stock. It’s tied more to the application and deployment side of the AI boom, which Dessai says should become more important as companies move from AI experimentation to real-world use.
Customers appear to agree: Innodata had a record-setting first quarter in 2026 with revenue up 54% year over year and adjusted EBITDA growth of around 96%. In one quarter, the company earned more revenue than in an entire year three years ago, according to company CEO Jack Abuhoff.
More from U.S. News
7 Best Defense Stocks to Buy Now
7 Best Semiconductor Stocks to Buy for 2026
7 Best Cybersecurity Stocks to Buy
6 Top Small-Cap AI Stocks and Emerging AI Companies originally appeared on usnews.com