Snowflake (SNOW 0.61%) captured the attention of many investors since its white-hot initial public offering (IPO) in 2020, and it continues to command attention even after the stock price's collapse since then. And for good reason.
Data storage, management, and analysis are all trends that aren't going away. Indeed, with generative AI in a hype cycle, the importance of digital data is only ratcheting up. All this helps explain why Snowflake just launched a new partnership with Nvidia (NVDA 1.98%), the semiconductor giant that emerged as a platform business for all things AI this year.
But there's another reason Snowflake needed to tap Nvidia for some new AI tooling.
Snowflake stock is still growing fast, but don't overpay
I recently wrote that as impressive as Snowflake's growth has been and is expected to be, most investors should approach this stock with caution. Snowflake's deal with Nvidia is proof.
Snowflake offers clients a treasure trove of data, and it gives businesses the tools they need to turn their own data into a treasure trove too. But Snowflake was still lacking the very specific toolset needed for custom generative AI model training (an example would be a big company making its own version of ChatGPT, trained using private information).
With this new partnership, Snowflake will leverage Nvidia's NeMo large-language model training algorithms, paired with Nvidia semiconductor systems, to allow Snowflake users to unlock the full potential of their data to train their own proprietary AI. It's a match made in heaven (or high up in the cloud).
But it also illustrates the fact Snowflake has work to do if it wants to continue growing at a rapid pace over the next few years. Remember, Snowflake still runs most of its workloads via Amazon's cloud infrastructure AWS, as well as on Microsoft Azure and Alphabet's Google Cloud -- all of which are also Snowflake competitors.
Another faster competitor, eyeing its own hot IPO?
Given its reliance on cloud giants' infrastructure and its need for Nvidia AI tools to offer its customers, perhaps Snowflake is not as in control of its own destiny as many investors think. And in addition to this, there's another smaller competitor making waves of its own: Databricks.
Databricks is a privately held and fast-growing business that offers data storage and management services like Snowflake. Databricks recently revealed that it grew its revenue by 60% in 2022 to over $1 billion for the first time. Rumors have it that Databricks is planning an IPO.
But one key difference between the two is that Databricks excels in AI. To bolster its services, Databricks reportedly just offered $1.3 billion to acquire the AI training start-up MosaicML, which competes with OpenAI, the creator of ChatGPT that Microsoft invested in.
With pressure from multiple sides and a new era of AI investment ramping up, there could be danger in overpaying for Snowflake stock right now.
The real stock to buy
Interestingly, there's one thing that AWS, Microsoft Azure, Google Cloud's infrastructure, OpenAI, Databrick's MosaicML, and now Snowflake all have in common: They all leverage Nvidia semiconductor systems and software, at least in part, for generative AI training. Indeed, Nvidia has become an incredibly powerful company, and the world is only just beginning to realize it.
That being said, don't rush to buy Nvidia either. The stock trades for a high premium to even next year's expected profitability. If you feel like you need to own either Nvidia stock or Snowflake stock, but are rightly worried about the high premium valuation, you might want to consider using a dollar-cost averaging plan to build a position over time.
Regardless of whether you buy either stock right now, Snowflake's new partnership with Nvidia reveals there are challenges ahead for Snowflake, and underscores the might of the new leader in the semiconductor and AI industry, Nvidia.
