A historic Snowflake Summit 2023

Snowflake Summit 2023 became the largest in-person Snowflake conference to date, with nearly 12,000 global in-person attendees in Las Vegas for four days. A total of 440+ sessions, 250+ customer speakers, 200+ ecosystem partners, and more were presented. You can hear some of the highlights from this year’s conference from our Senior Principal, Ejvind Hald:

1. LLM and Generative AI (announcement at Snowflake Summit 2023)

The hot topic at the conference was LLM and Generative AI. In a recently added Firechat, just prior to the conference, a new partnership with Nvidia and Snowflake was announced. This announcement came from the CEO at Snowflake, Frank Slootman, and the CEO at Nvidia, Jensen Huang. At some point, Snowflake customers will have access to GPU computational services directly from Snowflake. This is useful for many AI purposes, which require much computational power.

On top of this announcement, there were many sessions with LLM and Generative AI used in various contexts. It will be exciting to see what the future brings in this area.

2. Snowflake support for running apps in a container
(announcement at Snowflake Summit 2023)

This year’s surprise was the announcement of Snowflake Container Services, which will enable customers to run their own docker images inside the secure platform of Snowflake. The idea is to move apps to the data, and with this new existing feature, any of your favorite tools can be executed just next to the data in a secure way.

It will be quite interesting when released and enable seamless usage of existing apps within customer domains. Eventually, it will reduce complexity and save costs for the customer.

3. Dynamic Tables (preview)

Finally, the dynamic tables are now in public preview so all customers can use them. You simply declare the required contents of your table, and Snowflake will take care of the rest, including streaming pipelines and merging of new arriving content. These streaming pipelines can also be from Kafka, providing close to real-time content in Snowflake with configuration effort only.

This feature was announced at the Snowflake Summit 2022 and is now ready to use. It has some similarities to Databricks Delta Live Tables.

4. Streamlit in Snowflake (almost in preview)

The seamless integration of Streamlit in Snowflake Snowsight is almost ready to be released for preview to all Snowflake customers. It offers an easy-to-use code-based graphical representation of your data with many Streamlit objects available.

Recently, Streamlit has also added a data editor, so you can easily modify tables in Snowflake directly via a dashboard or similar. Streamlit in Snowflake is a powerful feature that offers simplicity and cost savings for customers.

5. Snowpark calculation for risks assessment for financial institutions (GA)

The Snowpark feature in Snowflake enables customers to write Python code directly to Snowflake and still enjoy high scalability and cost-efficient execution. One relevant use case for many financial institutions is stress and risk calculation models, and “Norges Bank Investment Management (NBIM)” demonstrated how they used these capabilities to create 100 million statistical calculations of scenarios in Snowflake in just 40 seconds!

In Intellishore, we can provide similar solutions, including seamless integration of Snowpark in dbt.