This actually depends a lot on how big the ELBs might scale. Each ELB could use up more than 100 IPs (Private & Public), so if they have 8 ELBs that all are taking the maximum amount of traffic they will need 800+ IPs.
When creating an ELB we check for 8 free IPs in each subnet to be used and won't let you create one without it. However, 8 is really the minimum we need for a single ELB to handle most workloads and failure cases safely.
Best practice is to give ELBs their own subnet(s), either by making one large (/18 - /21) one per AZ or by making a new one (like a /25 or /26) for each ELB with.
Charge for 77hs for attached elastic ipasked 3 years ago
ALB DNS name with 3 AZs enabled resolves to 1 IPasked 7 months ago
Do aws free-tier t3.micro servers have a max?asked 7 months ago
Why does ELB need one public IP address for each public subnet?asked 3 years ago
How to get FQDN of an instance without having an associated elastic IPasked 5 years ago
AWS Global Accelerator IP Subnet Range not up to date in ip-ranges.jsonasked 6 months ago
FSX - IP allocationasked 10 months ago
max. number of IP free ip per subnet for multiple ALBAccepted Answerasked 4 years ago
Transit Gateway - Multiple Subnet per AZasked 4 years ago
Private IP requirements in VPC for Application Load BalancerAccepted Answerasked 2 years ago