A few thoughts. 1) This is nothing surprising. People who build businesses based largely around filling in gaps in AWS's managed services hopefully expect and p

A few thoughts. 1) This is nothing surprising. People who build businesses based largely around filling in gaps in AWS's managed services hopefully expect and plan on AWS to attempt to fill that gap if that niche truly is profitable. 2) AWS is going to build what its customers ask for. I work for a relatively big consumer of AWS services, and we have a great relationship with our customer support team, who connect us to the engineering teams building the tools we use, and without fail down to the team, they are always very interested in how they can better serve our needs and build products we want to use. And if they hear "Kinesis is fine for some things, but we really want to also do Kafka" from enough people, they will build a Kafka managed service, just like they built ElasticSearch service, a Redis service, an RDS. And why shouldn't they? 3) For customers, AWS's cost model is far better than most managed cloud services. I don't know anything about Confluent's managed Kafka offering, but having everything metered in tiny increments is a huge reduction in the barrier to entry. If we want to build on a managed Kafka platform, but we want to start slow and grow as we need, it's so much better to be able to just spin up a tiny version to try things out, tear it down, take a break for other priorities, then come back with more confidence and more experience and iterate into building something real. Whereas too many cloud services require a huge lock-in agreement up front before you can get any support, or often any service. And it's not like the lock-in is necessary. These companies could be integrating their services to allow potential customers to leverage their AWS infrastructure. There's no reason these third parties can't use IAM credentials for authentication, and publish their services to customers over PrivateLink, a tool AWS built for explicitly this sort of purpose--to enable cloud providers to provide tight integrations with their shared customer base. 4) In terms of product quality, I don't think Confluent et al have anything to be worried about. AWS "competes" with Dropbox, technically, but is WorkDocs really serious competition? AWS's managed services have their place, but you can't use RDS to replace the flexibility and scale you can achieve by running PostgreSQL yourself. Their managed ElasticSearch and Redis products are okay, but ultimately very rigid, and nonsensical to use beyond mid-range scale. Their managed k8s service is something of a joke at this point (though I'm sure it will get better). If Confluent is truly offering a strong managed Kafka product, they will be just fine. 5) In fact, I suspect Confluent will get a lot more customers as a result of this announcement, as management at hundreds of mid-sized companies will now be convinced thanks to the AWS marketing that this Kafka thing their engineers have been begging for for years must be really key now, but they'll figure out how the AWS version doesn't quite live up and will end up being really expensive to scale to their company's needs. Then they'll look a little deeper and find Confluent. So I hope their marketing and sales teams are ready for the influx. It's coming.