So if an account is currently priced at $3 and a project envisions they will bring in 10,000 users initially. That's $30,000.

The code does not reflect this at all, as the barrier to entry is a variable number determined in real time by the witnesses.

We can already see that five out of 20 consensus witnesses are already signaling a reduction from 3 hive to 1.

I myself have concluded that it would be trivially easy to create off-chain solutions as well, such as a service that buys/sells/creates accounts and can issue them for "free" on credit, allowing users to interact with the chain for "free" while they slowly pay off the account creation fee. They would not have access to the owner key until the account was paid off. I think your assessment of this situation lacks the proper vision and only takes into account a static snapshot of the current situation we find ourselves in. You assume that the network will not react in the event that adoption increases. Rather, we will see dynamic changes occur in real time as that happens.