Now if VMFS collects the Oracle 8KB requests and puts them into 1MB blocks, then Im I correct to believe the SSD read cache pool will never be used?
No, your assumption is incorrect. The VMFS block size is solely for the sake of allocating/addressing files on the themselves. The VMFS block size has no influence of the IO issued by VMs.
That means an 8KB IO request of a VM will arrive as an identical 8KB IO request on the storage backend.
The VMFS block size is completely unrelated and transparent to both, the storage backend device and the VMs own IO behavior.