When everyone suddenly changes their minds, a certain kind of silence descends upon a market. For a large portion of the first part of 2026, artificial intelligence stocks were located there. The tone changed after two years of investors viewing anything with “AI” in the pitch deck as a lottery ticket. Money shifted from technology to consumer goods, pharmaceuticals, the safe, and the boring. During the first quarter, the Nasdaq fell into correction territory. And in the aftermath of all that selling, Morningstar’s analysts took an unconventional step: they went shopping.
There is a clear logic to their reasoning. Many AI-related stocks saw a selloff in the first quarter of 2026 due to a “anything-but-AI” sentiment following gains in 2025. Even as stocks have recovered, this retreat has left many well-capitalized names trading at a discount. The well-known behemoths Nvidia, Microsoft, Taiwan Semiconductor, Broadcom, and Meta remain on their list. However, the selection that consistently grabs my attention is the one that hardly ever makes headlines.
Snowflake would be that. It’s simple to pass. Unlike Nvidia, the company doesn’t produce chips, doesn’t have an eye-catching chatbot, and isn’t mentioned in earnings calls. It offers data infrastructure, which is the unglamorous plumbing that businesses require before any AI tool can be of any use. Strangely, that’s the entire point of contention. A disorganized pile of data cannot support the development of an intelligent application. It needs to be cleaned, stored, and arranged by someone. Wall Street seems to have forgotten to inquire about the location of all the data due to its obsession with hardware and foundation models.
Morningstar remembered. The company increased its fair value estimate on Snowflake, citing AI-related demand as the main factor behind the company’s recent success, after fourth-quarter results easily exceeded guidance and revenue growth accelerated to 30%. The current estimate is $223 per share. One figure hidden in the analysis really got my attention: almost 70% of Snowflake users are currently utilizing the AI features of the platform. That isn’t a pledge. That’s how people behave.
Naturally, the index is carried by the larger names. With a five-star rating and a stock Morningstar has described as “38% undervalued relative to its $600 fair value estimate,” Microsoft continues to be the company’s strongest conviction. For the majority of the year, investors penalized the company for its capital expenditures, just as they had previously penalized Amazon for constructing warehouses that no one could comprehend. It’s difficult to ignore the rhyme. Ten years ago, Tesla encountered similar skepticism, and its expenditures were viewed as reckless until they weren’t.
Taiwan Semiconductor occupies a more robust and quiet corner. Nearly every AI accelerator on the planet, including Nvidia’s, passes through Taiwan first because its factories produce about 70% of the world’s contract chips. The business selling jeans during the gold rush is the pick-and-shovel game.
This is all speculative. According to Morningstar, Snowflake has no economic moat, its valuation is high, and Databricks is pursuing the same clients. It’s possible that the thesis is incorrect. However, it seems like the audience sold the narrative and ignored the footnotes as the rotation went on. Usually, there is a reason why the stock is overlooked, but occasionally there isn’t.
