Valid question. I was rhe guy who embraced and optimized a full offering in our org, first on redshift and then on snowflake. Even merely evaluating a new db (on top of actually parsing their landing pages to understand what does what) for our data needs is exhausting to say the least. I still occassionally get enthusiastic with some new cool db (most recently materialize) and spend a weekend trying to spin it up and load our data, but I invariably walk away feeling redshift/snowflake is fine/better while simultaneously feeling guilty I didn't give the tool a real chance at all.

In the end, I'm probably going to give up on trying new db ideas unless their pitch absolutely blows me away. I've gotten much better ROI just doubling down on optimizing our code to use these popular offerings (which çontinue to add features every day) to our use cases and just throw a bit more money at it.

For example, we actually query a 115 billion row, 1.5 TB table in snowflake and barring some cluster warmup time, queries that return less than a million rows run in a couple hundred ms max. Until the day we need this to speed up to just a few ms, I'll probably stick here.

It's a really good question IMHO. The DB market has huge players that have been around for decades.

If you’re talking about time series databases then there are a number of them used in industrial settings called “process historians”, for instance, OSI PI.