Snowflake should emerge as a long-term artificial intelligence winner despite a host of near-term snowstorms, Wall Street analysts think. The cloud stock dropped more than 16% last Thursday after the company shared product revenue guidance that fell short of consensus expectations and results that indicated slowing growth. Wolfe Research downgraded Snowflake to peer perform from outperform, citing difficult macro overhangs. Even with these headwinds, many analysts remain positive on Snowflake’s long-term trajectory, viewing an acquisition and the transition to the cloud as two catalysts for the stock. SNOW YTD mountain Snowflake shares year to date “We view SNOW as one of the best positioned stories in infrastructure to benefit from generative AI tailwinds,” wrote Jefferies analyst Brent Thill in a May note to clients. He has a buy rating on Snowflake. He sees the acquisition of web search startup Neeva as another advantage enabling Snowflake to capitalize on the swarm of AI workloads moving to the cloud. Neeva’s AI software combining search with large language models is key to this thesis and should complement Snowflake’s aim to enable customers to integrate ChatGPT and chatbots into their platforms to find information — while also retaining privacy. The purchase should also boost Snowflake’s nontechnical user base and improve its conversation and search capabilities, said Piper Sandler’s Brent Bracelin, reiterating his overweight rating. “AI workloads often involve the processing of web-scale quantities of unstructured data,” wrote Raymond James analyst Simon Leopold, who has an outperform rating on Snowflake. “We are optimistic around the potential for incremental growth for Snowflake, though there exists substantial uncertainty around the ultimate size of the opportunity.” Snowflake’s Snowpark, which allows developers to write code in their preferred language, is also key to its AI reach. About 30% of customers used the platform on a weekly basis, with consumption improving 70% quarter over quarter, CEO Frank Slootman said during the company’s most recent earnings call. Analysts, including Goldman Sachs’ Kash Rangan, are looking ahead to Snowflake’s Summit conference in Las Vegas this month for further clues into its AI initiatives and progress. “We expect Snowflake Summit … to be a key catalyst for the stock as Snowflake updates LT financial targets and solidifies its AI product roadmap,” Rangan wrote. He has a buy rating on the stock. Deutsche Bank’s Brad Zelnick said in a recent note that AI, among other developments, should drive customer stickiness and improved use cases. “We reiterate our thesis that Snowflake sits in the sweet spot of two of the most important secular trends of this decade: data-driven decision-making (including AI) and public cloud adoption,” he said, maintaining his buy rating and $170 price target reflecting just 3% upside from Wednesday’s close. A murky future Not everyone seems optimistic about Snowflake’s AI potential, however. Redburn called the cloud company a “likely” loser in a May note to clients. The firm said the stock should not be viewed as a buying opportunity, citing its “inflated” multiple and steep valuation relative to companies such as Palo Alto Networks . “Hyperscalers have a clear path to driving reacceleration through ML/AI applications,” wrote analyst Alex Haissl in a May note. “However, for Snowflake, the deceleration and weak outlook stem partly from emerging structural hurdles, in our view.” SNOW 1M mountain Snowflake shares over the last month Some of those problems stem from declining spending among existing customers and challenges drawing new clients with deeper pockets. Haissl sees stronger growth opportunities for hyperscalers over Snowflake. While he views Snowflake as a top-notch user interface platform in a growing market, its status as a data warehouse creates limitations. “At its core, Snowflake is a data warehouse, a great but more mature technology with limited upside but also downside from ML/AI,” Haissl said. — CNBC’s Michael Bloom contributed reporting
