Re: ???
Not quite. SQL + ACID layers on Hadoop are ten a penny. Their main pitch was an OLAP layer, with cubes built transparently with Spark jobs and persisted into HBase for scalable/resilient/low latency retrieval.
Honestly it was really nice tech. Unfortunately it ran into the same problem almost every OLAP technology does: Almost nobody needs OLAP.
Especially in big data land there are almost no problems solved by OLAP that can't be solved (and arguably solved better) by a stonking great big fast MPP vanilla SQL - which every Hadoop cluster has 3-5 of kicking around already.
So the market was niche and the tech is still hard due to the curse of dimensionality and there are loads of established competitors. So they turned their hand to trying to do AutoML-style tooling and rebuilt all their kit to run on Kubernetes but at that point they were on iteration 3 and had no money and no credible product in a saturated market. So no rescue buyout/acquihire. So insolvency it is.