Let's say there's 0.1 bitcoin on an exchange. Someone bids $1M for it. That's $10M / bitcoin. If supply continues to shrink, for example, to 0.01. That's $100M / bitcoin. Infinite price is possible.
$100M per bitcoin is not infinite, and even if the supply shrinks to a single satoshi, which would imply a price of $10 trillion per bitcoin in your world, that is still not infinite.
Here's a more grounded example:
Let's say it's 1914 and there's only one known passenger pigeon left in the world. How much would someone pay for one more? I'll let you look up the answer, but I'll tell you now that it is far short of infinity.
Furthermore, now that there are 0 passenger pigeons in the world, what is the price of 1? Well, there are proposals to recreate passenger pigeons through genetic engineering and selective breeding. In your terms, supply is 0, and demand is positive, but do you really believe that someone will pay an infinite amount for that one passenger pigeon?
Your model is clearly broken.
I discuss a paradox, not a model. The current models of Bitcoin's price are broken. For example, the Stock-to-flow model. It runs into this paradox. S2F is a model that references itself. It has no grounded reality. I think it's broken. If Bitcoin's price is about supply and demand, any model is a self-referencing model.
I read the pigeon story. The pigeon is worth as much as someone is willing to pay for it. The same logic applies to Bitcoin. Bitcoin is worth as much as someone is willing to pay for it. How should we determine the price? Bitcoin uses Proof of Work mining. This translates to how much energy we spend on mining Bitcoin. Is it 10%, 20%, 50%, 80% of the world's energy? The paradox is the energy is limited and we can't determine Bitcoin's price.