The AWS documents clarify this. When you get 1 vCPU in a Lambda you're only going to get up to 50% of the cycles. It improves as you move up the RAM:CPU tree but it's never the case that you get 100% of the vCPU cycles.
However that's not due to aws overhead or oversubscription but x86 architecture. For production workloads 2 vcpu should be minimum recommendation.
On ARM where 1 vcpu = 1 core it is more straightforward.
The AWS documents clarify this. When you get 1 vCPU in a Lambda you're only going to get up to 50% of the cycles. It improves as you move up the RAM:CPU tree but it's never the case that you get 100% of the vCPU cycles.