Artificial intelligence is creating investment opportunities in the less visible work needed to make applications useful. Models require specialized training data, businesses need reliable recovery and security, and developers must manage the cost of running software that calls large language models. At the device level, AI also needs processors that can handle images and other inputs without sending every task to a remote data center. These markets provide several ways to examine AI adoption beyond the largest platform companies, though narrower businesses can bring greater customer concentration and execution risk.
Innodata illustrates the importance of the data layer. Its August 6 release reported second-quarter revenue of $92.1 million, up 58% year over year, and described new work involving reinforcement learning for long-horizon and computer-use agents. Yet two customers represented 37% and 34% of quarterly revenue. The company also said its June cash balance included customer prepayments for pass-through costs. Expanding demand for data services therefore needs to be evaluated alongside customer dependence and the distinction between cash received and cash freely available to shareholders.
Cyber resilience is another part of the same ecosystem. Rubrik reported fiscal second-quarter 2027 revenue of $427.3 million, up 38%, and launched Rubrik AI across its security and agent offerings. Its release described autonomous action with guardrails and recovery workflows. That opportunity comes with a financial trade-off: reported free cash flow was $65.7 million, compared with $57.5 million a year earlier, while the company still recorded a GAAP loss per share. More revenue and a broader product range do not automatically settle the question of profitable scaling.
The applications themselves can also provide direct commercial exposure. Duolingo Max offers generative-AI features such as Video Call and Roleplay within a paid subscription tier. Fastly’s AI Accelerator addresses a different problem by caching semantically similar model responses, potentially reducing repeated requests and inference costs. These examples show that an AI business need not own a frontier model. Its opportunity may come from packaging a useful customer experience or making another company’s model cheaper and more practical to use.
Physical devices create a further set of possibilities. Ambarella’s September 3 results described record edge-AI revenue and growth from its CV75 and CV72 chips. The company supplies an AI platform across applications including cameras, vehicle safety and robotics. Meanwhile, IonQ completed its acquisition of semiconductor foundry SkyWater on July 31. SkyWater has described packaging capabilities intended for AI accelerators and high-performance computing. That gives IonQ a separate AI-related infrastructure connection; it does not establish that its quantum-computing research has already become a mature commercial AI business.
Each route has a different way of succeeding and a different way of disappointing investors. Data suppliers must retain demanding customers, security vendors must convert adoption into cash, and chip suppliers must execute product ramps while controlling costs. Businesses with several activities also require care when attributing growth to AI. A strong company-wide sales increase may reflect acquisitions, non-AI products or a small comparison base. The useful question is whether the commercial opportunity can produce durable returns at the price investors pay, rather than whether the company belongs to a fashionable theme.
Zapp2Photo/Shutterstock.com
Our Methodology:
Before collecting financial data, we fixed a 73-stock coverage universe. We required a current U.S.-listed operating company, primary-source documentation of a direct commercial AI product, service or AI-related semiconductor or data-center supply activity, positive trailing revenue, and at least 10% trailing-12-month revenue growth. We limited eligibility to companies held by 50 or fewer hedge funds in Insider Monkey’s Q2 2026 database of 1,006 funds. That threshold is our defined lower-ownership proxy for “overlooked”; it does not measure public awareness or prove that any stock is undiscovered.
Using Stock Analysis income statements accessed on October 3, 2026, we ranked eligible companies by reported whole-company trailing-12-month revenue growth, highest first, and selected 11. Ties would be resolved by greater hedge-fund ownership, then alphabetical ticker order. We imposed no profitability requirement and excluded missing data, companies below the growth threshold, and businesses without verified direct commercial AI exposure. Quantum-machine-learning research alone was insufficient. This defined coverage screen is not an exhaustive search of all AI-related companies.
We display revenue growth and the fiscal period ending the trailing window for every ranked company. These growth figures are not AI-segment growth, organic growth or forecasts; acquisitions and small comparison bases can affect them. IonQ qualifies through the current AI-related packaging activities of its acquired SkyWater subsidiary. Its June 2026 trailing figures predate the July 31 acquisition, and therefore describe its prior business mix rather than SkyWater’s contribution. Quarter-end hedge-fund counts also predate that acquisition. The list counts down from 11 to 1, and “to buy” identifies candidates for further valuation work rather than a promise of investment returns.
11. Dynatrace, Inc. (NYSE:DT)
Number of Hedge Fund Holders: 48
TTM Revenue Growth: 17.94%
Fiscal Period End: June 2026
READ ALSO: Is Dynatrace, Inc. (DT) A Good Stock To Buy Now?
10. Ambarella, Inc. (NASDAQ:AMBA)
Number of Hedge Fund Holders: 33
TTM Revenue Growth: 20.04%
Fiscal Period End: July 2026
9. Fastly, Inc. (NASDAQ:FSLY)
Number of Hedge Fund Holders: 28
TTM Revenue Growth: 20.35%
Fiscal Period End: June 2026
READ ALSO: Fastly (FSLY) Is Getting More Interesting. Meta’s Muse Is Part of the Reason
8. SentinelOne, Inc. (NYSE:S)
Number of Hedge Fund Holders: 41
TTM Revenue Growth: 21.09%
Fiscal Period End: July 2026
7. GitLab Inc. (NASDAQ:GTLB)
Number of Hedge Fund Holders: 50
TTM Revenue Growth: 22.99%
Fiscal Period End: July 2026
6. Duolingo, Inc. (NASDAQ:DUOL)
Number of Hedge Fund Holders: 39
TTM Revenue Growth: 29.36%
Fiscal Period End: June 2026
READ ALSO: Duolingo (DUOL) vs Coursera (COUR): Which is a Better Stock to Buy?
5. Innodata Inc. (NASDAQ:INOD)
Number of Hedge Fund Holders: 25
TTM Revenue Growth: 39.02%
Fiscal Period End: June 2026
READ ALSO: Innodata (INOD) Is No Longer Just a Service Provider And Analysts Are Taking Notice
4. Rubrik, Inc. (NYSE:RBRK)
Number of Hedge Fund Holders: 44
TTM Revenue Growth: 42.45%
Fiscal Period End: July 2026
3. Upstart Holdings, Inc. (NASDAQ:UPST)
Number of Hedge Fund Holders: 36
TTM Revenue Growth: 45.33%
Fiscal Period End: June 2026
READ ALSO: Why Upstart (UPST) Still Has an AI-Lending Moonshot Case
2. SoundHound AI, Inc. (NASDAQ:SOUN)
Number of Hedge Fund Holders: 20
TTM Revenue Growth: 54.58%
Fiscal Period End: June 2026
1. IonQ, Inc. (NYSE:IONQ)
Number of Hedge Fund Holders: 42
TTM Revenue Growth: 370.64%
Fiscal Period End: June 2026
READ ALSO: IonQ vs. Rigetti: Is the Quantum Rally Backed by Fundamentals, or Just Momentum?
READ NEXT: SoundHound AI Closed Its LivePerson Deal. Does the Acquisition Fix SOUN’s Scale Problem or Import LPSN’s Weaknesses? and Ambarella Beat Estimates but Fell as NXP Buyout Talk Lingered. Is the Edge-AI Upside Already Priced In?
