Snowflake new directions

A few points on Snowflake. Starting a new thread as the last post took a veer from the (well deserved) love letters to Saul.

Also, if you are new to the story, make sure you go back and re-read Smorg’s thread on Snowflake, incl the second post on competition and data sharing.…


I also completely agree w/ Smorg’s earnings take on the intentional short term impact for long term gain:…

Good stuff. I will add that factoring back in the $160M direct impact from these improvements into FY guidance gives an 80.6% guide — right under the last FY guidance a year ago. This company is slowing down because of its own performance improvements.

It’s up to you if that is acceptable or not (is it an improvement or impact?), but I like it a lot. A 10-20% improvement to customer pricing for compute (with impact to their revenue) all for the benefit of customers and their long-term durability of the platform, plus their new insight that it leads to more customer workloads moving towards the platform. (Eventually filling in 35-40% of that loss, per their estimate.)


All these improvements to the programmability have been greatly enhancing the data science and data engineering (data movement) types of workloads.

There were strong hints in the CEO intro that they are moving into “new workloads” – which to me, means they are attacking new markets and audiences. The current workloads are for data warehousing, data engineers, data lakes, and data science. I think we’ll see some interesting announcements at Snowflake Summit in June, and into the fall. But I think part of this is trying to take over existing data lake workloads in a more centralized tool.

They gave later hints of what some of those might be… but I think the hint might have been misinterpreted in the last post, which said: Snowflake is also moving into security (like DDOG).

Snowflake is not moving DIRECTLY into cybersecurity, such as where Datadog is going. Snowflake was pointing out that they are a good platform for analyzing a large quantity of data – such as a security-focused data lake. Here is the quote from the CEO in response, with the key word (IMHO) highlighted: We in Snowflake are just an ideal platform for hosting that type of capability. So, you will see us lean into that opportunity a lot more going forward.

I look forward to seeing what Snowflake comes up with here, but it won’t be directly becoming a cybersecurity company. (And I am curious about what Frank was talking about the “add-on motion” was he was referring about. Ability to analyze external data lakes?).

Lacework is one such cloud security company that has been built upon Snowflake. Observability is another such industry that needs to analyze a lot of data (in this case metrics and logs), and there again, we have Observe as a company that has built upon Snowflake as a platform.

I would not assume Snowflake is entering these markets directly. They are focused most on adding new types of workloads that their customers want to run in their centralized Data Cloud.


One of the exciting pieces of news from their earnings was their acquisition of Streamlit, an open-source framework for creating, deploying and sharing data apps. Data scientists can use this to create web apps like interactive dashboards over a set of data, but without having to be a Javascript or front-end developer. They can do it all in a simple Python script.

It was $800M (80% stock) for this basically revenue-less company, that just started an enterprise managed service for hosting Streamlit in Nov-21.

This adds a new dashboarding capability that combines extremely well with their new Snowpark for Python. Data scientists can now start to use a Spark-like programming interface for data manipulations, and now, with an embedded Streamlit capability, can easily create new web apps that run within the Data Cloud over that data.

Exciting acquisition.

  • muji

Hi Muji

re Snowflake use cases in security - could it be that their database structure is just much more capable in analyzing massive amounts of data in real time - faster than anything else out there right now? Speed and flexibility is what they seems to excel at.

1 Like