I see.

Prior to joining Snowflake, I had proposed and received approval from Andy Jassy to purchase Snowflake while I was at AWS in 2017. However, the deal fell through due to internal politics at AWS, particularly opposition from the Redshift team.

Where does that figure come from? When you listen to Schluteman, Jassy, Solipsky, and others, they say only about 10% of the workloads are on the cloud.

Well, that depends on your definition. At the infrastructure level, it's in the high 40s. However, a lot of the software is still legacy software that's just been moved to the cloud.

Understood. Why is Snowflake benefiting in 2027?

If we examine where cloud providers generate their earnings and revenues, about 70% of revenue comes from foundational infrastructure services like compute, storage, and network. These, however, have relatively low margins, usually in the 20% to 30% range. The reason for this is Amazon's pricing, which everyone else has to match. If Amazon didn't exist, we'd still have a cloud, albeit smaller and more expensive, and probably under a different name.