Many artificial intelligence (AI) stocks skyrocketed in value in recent years as more companies realized they could accelerate and automate their operations with large language models (LLMs) and generative AI applications. The obvious winners include Nvidia, which produces the data center GPUs for processing those AI tasks, and Microsoft, which acquired a big stake in ChatGPT's creator OpenAI and integrated its AI tools into its own cloud services.
While Nvidia and Microsoft are still great long-term AI plays, investors shouldn't overlook the other under-the-radar AI plays that could generate big gains over the next few years. Three of those underappreciated AI-oriented stocks are Duolingo (DUOL -0.35%), Confluent (CFLT 1.37%), and MongoDB (MDB 0.76%). Let's look at how these three lesser-known AI stocks could make you a fortune.
Image source: Getty Images.
Duolingo
Duolingo, which owns the world's most downloaded language learning app, might not seem like an AI company. But it's been using generative AI to produce its online courses at a faster rate, replacing a lot of its human contractors with AI-powered services, and expanding its premium Duolingo Max tier with AI-driven conversations. It also uses its own "Birdbrain" LLM to customize its lessons for its users based on their proficiency, pace, and learning styles.
Duolingo served 130.2 million monthly active users (MAUs), 46.6 million daily active users (DAUs), and 10.3 million paid subscribers in the first quarter of 2025. That's up from just 40.5 million MAUs, 9.6 million DAUs, and 2.5 million paid subscribers at the end of 2021.
From 2024 to 2027, analysts expect Duolingo's revenue and EPS to grow at a CAGR of 29% and 51%, respectively. That growth should be driven by the expansion of Duolingo Max's AI-driven services, more non-language subjects like math and music, new pricing tiers, and stickier gamification features (like leagues and streaks). Duolingo's stock isn't cheap at 115 times next year's earnings, but its rapid growth, its dominance of the language learning market, and its expanding AI ecosystem all justify that higher valuation.
Confluent
Confluent's cloud-based platform processes "data in motion" as it flows between different applications within an organization. Its namesake platform runs on Apache Kafka, an open-source platform for processing streaming data, but it integrates additional analytics services to differentiate itself from other "Kafka-as-a-service" providers.
Confluent doesn't build its own large language models or generative AI platforms, but it accelerates the delivery of data for those services. That's why it set an ambitious goal for becoming the "central nervous system for modern digital enterprises" in its IPO filing four years ago. Its total number of customers grew from 3,470 in 2021 to 6,140 in the first quarter of 2025, and the market's demand for its streaming data services should continue rising as the AI market grows.
From 2024 to 2027, analysts expect Confluent's revenue to rise at a CAGR of 19% as it narrows its net losses. Its near-term catalysts include its deeper partnerships with Microsoft's Azure, Alphabet's Google Cloud, OpenAI, Databricks, and other cloud and AI leaders, its growth among larger enterprise customers, and the expansion of its ecosystem. Confluent's stock still seems reasonably valued at 6 times next year's sales, and it should still have plenty of room to grow as the rapid expansion of the AI market drives companies to stream more data across its digital pipelines.
MongoDB
MongoDB is another company that helps companies organize large amounts of data for AI applications. Its namesake platform allows companies to store large amounts of unstructured data in a non-relational database. That approach differentiates it from older relational databases, which only store their data in rigidly structured tables and rows, and makes it easier for its clients to customize their data for specific tasks.
MongoDB's subscription-based cloud-based service, Atlas, allows its clients to analyze all of that data. Its MongoDB Copilot -- a generative AI assistant that was launched last year -- streamlines, optimizes, and accelerates those queries. It also uses AI to detect suspicious and misconfigured patterns within its database. It served 57,100 customers in the first quarter of fiscal 2026 (which ended this April), up from its 33,000 customers at the end of fiscal 2022.
From fiscal 2025 to fiscal 2028, analysts expect its revenue to grow at a CAGR of 16%. It's not profitable yet and might not seem cheap at 7 times next year's sales, but its expansion of Atlas, its recent acquisition of Voyage AI (which adds embedding and reranking models to its apps), and new AI partnerships should fuel its long-term growth.
