Why are investment funds led by Warren Buffett and Brad Gerstner dumping this artificial intelligence (AI) stock?

Technology stock Nvidia the price is up 151% so far in 2024. Meanwhile, the tech stock Snowflake (SNOW -3.06%) it trades 40% over the same period. As it turns out, having strong connections to the artificial intelligence (AI) sector is not enough on its own to create value. Investors also expect to see solid financial results.

Snowflake is a provider of cloud computing software solutions and generated explosive revenue growth when its shares first went public in 2020. That growth has since slowed as the company continues to spend heavily on developing new AI products and services at the same time. time. Increased expenses cause burst losses at the bottom line. It also made institutional investors rethink their investment thesis in Snowflake.

Top mutual funds are heading for the exits

of Warren Buffett Berkshire Hathaway The holding company bought Snowflake shares ahead of its initial public offering in 2020. Although the price it paid was not made public, Berkshire likely paid about $120 per share. It drove the stock to an all-time high of $392 at the end of 2021, but continued to hold on the way back down. Berkshire eventually sold its entire Snowflake stake in the second quarter of 2024, when it was trading at about $135 per share.

Prolific tech investor Brad Gerstner led Snowflake’s Series C investment round in 2015 (when it was still a private company) through his fund, Altimeter Capital, so he was an early believer in its potential. The fund held 15.4 million shares of Snowflake in the third quarter of 2023, but has since trimmed that position by 38%. It sold nearly 3 million shares in the last quarter of 2023, followed by another 2.1 million shares in the first quarter of 2024 and another 730,343 shares in the last second quarter.

Why did Buffett and Gerstner apparently turn bearish?

Snowflake is gearing up for the AI ​​revolution

Large and complex organizations using multiple cloud providers often find that their data is highly fragmented as it is spread across different platforms. Snowflake’s Data Cloud is a revolutionary tool that can bring it together in a platform-agnostic way so that the organization can more easily analyze it and extract valuable information that can be used to improve day-to-day operations.

Last year, Snowflake made the leap into AI products and services when it launched Cortex AI, which is designed to complement the Data Cloud. The Cortex platform allows companies to combine their data with ready-made large linguistic models (LLMs) such as Meta platformsLlama 3 and Mistral Large so they can build AI software applications.

Cortex also comes with several pre-built AI tools. Cortex Search uses natural language processing so developers can instantly retrieve data from large sets with a simple prompt. Then there’s Document AI, which can quickly extract valuable information from unstructured sources like contracts and invoices. In the past, this type of data could only be processed if a human worker manually read each document and transferred the information into a usable format.

At the end of Snowflake’s second fiscal quarter 2025 (ended July 31), approximately 2,500 of its 10,249 customers were using its AI products on a weekly basis. This suggests solid uptake given that Cortex was launched only a year ago.

Image source: Getty Images.

Snowflake’s revenue growth is decelerating

While everything highlighted above sounds great, AI isn’t necessarily driving Snowflake’s bottom line just yet. The company generated $829.3 million in product revenue during the second quarter, an increase of 30% year over year. That’s a strong growth rate at face value, but it marked a significant deceleration from the 37% revenue growth Snowflake delivered in fiscal 2Q24 and the 83% number it delivered in fiscal 2Q20 2023.

Snowflake spent a record $936 million on operating expenses during the second quarter, which was a 26 percent increase from the year-ago period. Its research and development expenses — which are used to develop new products — reached $437 million for the quarter, a 40 percent increase.

In other words, Snowflake experienced a slowdown in revenue growth despite spending more money on growth-oriented initiatives, which is not a great sign. It also had major consequences for the company’s bottom line, with its net loss amounting to $316.9 million for the quarter, a 40% increase over the same quarter last year.

There is some evidence that the spending will pay off going forward, as Snowflake’s remaining performance obligations (RPOs) rose 48% to $5.2 billion in Q2. The company expects to turn half of its backlog into revenue over the next 12 months, but isn’t providing guidance for the rest — so while it’s possible accelerated revenue growth could follow, it is not a certainty.

Snowflake’s rating is another downside

Despite Snowflake stock’s 71% drop from its all-time high, it can be argued that it’s still quite expensive. Based on the company’s trailing 12-month product revenue of $3.1 billion and its market capitalization of $38.2 billion, its stock trades at a price-to-sales (P/S) ratio of 12.

It is difficult to compare Snowflake to other publicly traded companies due to its unique product portfolio. However, Microsoft, Amazonand Alphabet are the three largest cloud service providers in the world and also happen to be leaders in the AI ​​space.

Based on the P/S ratio, Snowflake is slightly cheaper than Microsoft, but it is much more expensive than Amazon and Alphabet:

PS report data by YCharts

I think Snowflake’s valuation is still too high on this basis. All three tech giants have decades-long track records and are each highly profitable, while Snowflake continues to burn through significant cash.

In my opinion, Snowflake stock has never been a good fit for Berkshire’s portfolio. Buffett likes to invest in companies with steady growth and robust profitability because they often have the flexibility to return money to shareholders through dividends and share buybacks. Snowflake has a buyback program, but it may not be sustainable given the company’s mounting losses, and that likely rules out dividend payments in the near future.

Snowflake is in the wheelhouse of Gerstner and his Altimeter fund, however, because it focuses exclusively on technology stocks. That’s probably why he still owns 9.2 million shares in the company. He added to his holdings in stocks such as Alphabet, Amazon and Uber technologies in the first half of this year, and since it may have better short-term prospects than Snowflake, don’t be surprised if it continues to sell it to free up capital for those opportunities.