Warren Buffett, the renowned CEO of Berkshire Hathaway, oversees a massive stock portfolio valued at $287 billion, in addition to numerous private subsidiaries.
Buffett’s simple investment strategy has consistently delivered market-beating returns for nearly six decades. He typically invests in companies demonstrating steady growth, reliable profitability, and strong management. Companies with shareholder-friendly approaches, such as dividend payouts and stock buyback programs, are particularly attractive to him. Surprisingly, even with the rise of artificial intelligence (AI), Buffett has avoided chasing the latest stock market trends.
Berkshire invested in a cloud computing firm called Snowflake in 2020, a move that appeared somewhat unusual given Buffett’s typical investment criteria. Consequently, many weren’t shocked when Berkshire sold its entire position last year. However, since the sale, Snowflake’s stock has gained roughly 30%. Did Buffett and his team exit too early, and could this be a good long-term buying opportunity for investors?
Snowflake’s AI Ambitions
Snowflake’s data cloud allows businesses to consolidate their data onto a single platform, enabling more effective analysis and the extraction of valuable insights. This tool is particularly beneficial for large, complex organizations that utilize multiple cloud providers, which often results in fragmented data silos. As data serves as the essential element for most AI software, Snowflake is becoming a key platform for developing and deploying this technology.
In late 2023, the company introduced the Cortex AI platform, which provides businesses access to large language models (LLMs) from prominent third parties like Anthropic, OpenAI, DeepSeek, and Meta Platforms. Businesses can integrate their internal data with these LLMs to develop customized AI software. In 2024, Snowflake extended the platform with the Cortex Agents, which can create virtual assistants capable of handling specialized tasks.
For example, a business could design an intelligent sales assistant capable of analyzing various types of data, including conversation transcripts between sales staff and clients, to quickly source key information. This feature can significantly decrease the time employees spend manually searching documents.
At the end of fiscal 2025 (January 31), Snowflake had 11,159 customers, with the company reporting that over 4,000 of them were already utilizing the company’s AI products on a weekly basis.
Decelerating Revenue Growth and Mounting Losses
When Berkshire invested in Snowflake before its initial public offering (IPO) in 2020, the company was consistently growing its annual revenue at a rate of over 100%. Although it is unrealistic for any company to maintain such rapid growth indefinitely, Snowflake’s top-line expansion has notably decelerated, despite increased spending on customer acquisition and new product development.
During fiscal 2025, Snowflake generated a record $3.4 billion in product revenue, but experienced a growth rate of 30%, marking its slowest pace as a public company. The company’s operating costs rose by 28.8% to reach $3.8 billion, which included a 38.5% increase in research and development spending to support its expansion into AI. These higher costs led to a substantial net loss of $1.3 billion, a 53.7% increase compared to the prior year. The company is betting that its significant investments in AI will eventually pay off.
Early indicators suggest potential success, with Snowflake’s remaining performance obligations (RPOs) increasing by 32.6% year over year during the fourth quarter of fiscal 2025, reaching a record $6.8 billion. RPOs are like a backlog of orders, indicating strong demand for the company’s services. However, Snowflake expects to convert only 48% of them into actual revenue in the next 12 months. Furthermore, management forecasts that product revenue will be $4.2 billion in fiscal 2026, reflecting a growth of just 24%.
Snowflake’s Valuation
While the exact price Berkshire paid is unknown, Snowflake went public at $120 per share. Berkshire divested its entire position in the second quarter of 2024 (ending June 30, 2024), with the stock trading at around $135 at the end of the quarter. As a result, Berkshire likely earned a return of about 12.5% during its four-year holding period. This return is not particularly impressive, when compared to Berkshire’s historical average. Snowflake stock, therefore, was a drag on Berkshire’s overall performance.
Since Berkshire’s sale, Snowflake stock climbed by 30%. However, given its current valuation and the company’s slower growth and substantial losses, the stock’s current price is hard to justify.
Snowflake trades at a price-to-sales (P/S) ratio of 16.2, making it considerably more expensive than other cloud and AI leaders such as Microsoft, Alphabet, and Amazon.
It is unlikely that Buffett personally decided to add Snowflake to Berkshire’s portfolio. He generally sticks to businesses he fully understands, and early-stage technology companies typically fall outside his areas of expertise. One of Buffett’s lieutenants likely made the purchase. Given Snowflake’s valuation and its position at the time of the sale, the decision to sell was not surprising. Missing out on potential future gains doesn’t necessarily mean the sale was a mistake. Investors can only make decisions based on currently available information, and there was no compelling fundamental reason for Snowflake to continue climbing, given the challenges it faced. Therefore, investors may want to consider following Berkshire’s lead and staying on the sidelines when it comes to Snowflake stock.
