(B) The recent surreal time, due to Covid 19, has been a challenge for many start-ups in Silicon Valley since the capital markets have been dried for new IPOs. Without any liquidity events, start-ups usually struggle to raise capitals and grow. But this has changed recently as many companies (Palantir, Unity, Airbnb…) have recently or are planning an IPO…But out of all those IPOs, the one that I am really excited about is SnowFlake!

SnowFlake helps large enterprise companies to store and analyze billions of user, product, supply chain, and/or any enterprise records. Many Fortune 500 companies are moving their legacy data warehouses from IBM/Netezza, Microsoft, Oracle, and Teradata to SnowFlake. I do not have any statistics to support this premise, but from my own experience, and people from my own network who work in big data, everyone seems either to have already migrated or planning to migrate its data warehouse to SnowFlake.

SnowFlake has three key competitive advantages:

1. I believe that more and more Fortune 500 companies will have two cloud providers in the future. Most of Fortune 500 companies have only one cloud provider today: AWS, Microsoft, or Google but will be moving to have two providers to limit their dependancies and lower their risks with one provider. And, SnowFlake data warehouse works on any cloud: AWS, Microsoft, or Google. A company that uses both AWS and Google cloud might be more incline to use SnowFlake that works both on AWS or Google than having all its data into AWS Redshift or Google Bigtable.
2. SnowFlake has a neat architecture: it uses virtual compute instances for its compute needs, and a storage service for persistent storage of data. And has many useful features from time travel, support for SQL queries that access semi-structured data, to connectors for many types of clients…
3. SnowFlake has built a strong ecosystem and many tools for developing analytics. A good use case is for instance the possibility to develop ETLs where the “E” of the ETL e.g. the extraction of the data can be done in SnowFlake, and the “T” e.g. the transformation of the data can be done in a big processing data engine such as Spark. You read data from SnowFlake SQLs in a Spark job, and you can write back data from a Spark job to SnowFlake.

So what could go wrong?

1. As it is always the case, platform providers such as AWS, Microsoft, and Google can compete more aggressively on price, and offer better integration of their data warehouse with their other cloud services. For instance, Amazon Redshift is well integrated with Amazon EMR, Amazon Athena, Amazon Kinesis, and Amazon SageMaker services.

Note: The picture above is the main building of SnowFlake close to highway 92 in San Mateo.

Copyright © 2005-2020 by Serge-Paul Carrasco. All rights reserved. Contact Us: asvinsider at gmail dot com.