Thread by @QuinnyPig

AWS is committed to customer success, but if the customer grows too successful then apparently it's time to pivot to compete more directly with them. cnbc.com/2020/09/16/sno…

So, @SnowflakeDB as per this article / their S-1 has committed to spending $1.2 billion on @awscloud. That's large but not ridiculous at the "giant customer" scale given the timeframe.

I'd see Snowflake's increased cloud service usage from @Azure and @GCPcloud as being less about "reliance on @awscloud" and more about "customers are in those clouds, and don't want to pay data egress fees for an entire data warehouse."

I saw a take earlier that @SnowflakeDB's marginal cost for additional users is "basically nothing."

I assure you that is not true. To wit:

AWS's RedShift was originally aimed in 2012 at taking workloads from Oracle. Then Snowflake launched two years later and started winning RedShift business away.

After a slumber, a LOT of RedShift features came out last year.

As mentioned periodically, one of the things I do is help negotiate large scale commitments / contracts for @awscloud customers. I have some thoughts on what this article says next.

duckbillgroup.com/services/aws-c…

I've yet to see a negotiation where AWS salesfolk pushed a customer explicitly to migrate a workload from a third party vendor into an AWS native offering. I get that @awscloud signs a lot more enterprise contracts than I have customers, but still...

This paragraph rings true, but only if you replace "sales team" with "team focused on figuring out how RedShift wasn't as capable." That would track with the 2019 raft of RedShift improvements.

While RedShift's advanced query accelerator was great, I preferred the ra3 instances for tiering data to S3 and the ability to put instances to sleep at night so you weren't paying for idle.

This is fascinating and a welcome sign, but it's occluded by the sad fact that @SnowflakeDB CEO Frank Slootman mispluralized "field staves."

I think @awscloud's kneejerk reaction of "compete with SnowFlake" is ill founded *if true*. The data is going to live in AWS either way; does AWS really care all that much whether it's in SnowFlake's account or the end customer's? (They might, based upon discounting!)

The problem here is that basically nobody is going to read an article like this, stand up, and say "@awscloud competing with its customer? NONSENSE!"

Their reputation for poor partnership haunts them, and badly needs a reset. Reputations take a long time to shift.