Thanks for sharing your video and modeling ideas.
It is not clear why your regression curve is curved on the log scale. I would explore a long term equilibrium model like this.
logPrice = a + b*Date + c*logRev + error
If you assume c=0, this model is linear on log price scale. The inclusion of the logRev term makes the explicit assumption that Price and Revenue are cointegrated, that is their ratio stays close to some mean value. Another way to think this model is that is like saying that Price and Revenue are related to each other (via a power law) with a steady shift over time.