What causes WACC to invert?

Hi together, I'm currently working on a valuation and cannot figure out my mistake (if it is one). I calculate WACC by using the basic formula:

WACC = (E/V) * Cost of Equity + (1-(E/V)) * (1-t) * Cost of Debt

Since the company is private I calculate the Cost of Equity using the CAPM and a relevered beta. The formula therefore becomes:

WACC = (E/V) * (rf + BETAu * (1 + (1-(E/V))/(E/V)) * (1-t)) * MRP) + (1-(E/V)) * (1-t) * Cost of Debt

I noticed with my specific inputs (infrastructure company = very low unlevered beta, low risk free rate, low cost of debt) WACC declines when I increase (E/V) which seemed counterintuitive. I just differentiated the above formula (by hand so there might be a mistake) with respect to (E/V) and got the following result:

dWACC/d(E/V) = rf + t * BETAu * MRP - (1-t) * Cost of Debt

Looking solely on the deviated formula I can see that if the last part of the equation (= (1-t) * Cost of Debt) becomes greater than the first part (= rf + t * BETAu * MRP) WACC should decrease for an increase in (E/V). Mathematically understood (correct me if my maths has errors) but how can expain the result economically?

I most likely expect that indeed the math is wrong. The effect can happen when the asset yields an (unlevered) return that is lower than the cost of debt (which also seems counterintuitive in this case)

I would be grateful for any help!