It looks like the author specifically talks about the infra for an early-stage startup that has not found product-market fit yet. If a startup has product for consumers and does find the product-market fit, then I'd imagine two pieces of infrastructure that is hard to come by: EC2, and S3. Yes, EC2, the grandpa's infra that people either ignore or despise. But really, anyone can learn how to set up and run a k8s cluster, yet very few companies can offer something like EC2: full abstraction of the underlying servers, worry-free of provisioning new servers, and robust and versatile implementation of dynamic autoscaling. After all, all the k8s shit won't scale easily if we can't scale the underlying servers.
And S3. S3 is just a wonderful beast. It's just so hard to get something that is so cheap yet offers practically unlimited bandwidth and worry-free durability. I'd venture to say that it's so successful that we don't have an open-source alternative that matches S3. By that I specifically mean that no open-source solution can truly take advantage of scale: adding a machine will make the entire system more performant, more resilient, more reliable, and cheaper per unit cost. HDFS can't do that because of its limitation on name nodes. Ceph can't do that because of it bottleneck on managing OSD metadata and RGW indices. MinIO can't do that because their hash-based data placement simply can't scale indefinitely, let alone ListObjects and GetObjects will have poll all the servers. StorJ can't do that because their satellite and metadata servers are still the bottleneck, and the list can go on.
And S3. S3 is just a wonderful beast. It's just so hard to get something that is so cheap yet offers practically unlimited bandwidth and worry-free durability. I'd venture to say that it's so successful that we don't have an open-source alternative that matches S3. By that I specifically mean that no open-source solution can truly take advantage of scale: adding a machine will make the entire system more performant, more resilient, more reliable, and cheaper per unit cost. HDFS can't do that because of its limitation on name nodes. Ceph can't do that because of it bottleneck on managing OSD metadata and RGW indices. MinIO can't do that because their hash-based data placement simply can't scale indefinitely, let alone ListObjects and GetObjects will have poll all the servers. StorJ can't do that because their satellite and metadata servers are still the bottleneck, and the list can go on.