Understanding SOFR Swap Spreads

Have been self-studying rates and must admit I am stumped by the existence of SOFR swap spreads. I am stuck between two conflicting ways of thinking about them; the first supports their existence, while the second suggests they should not exist. I've detailed my thoughts below. Would greatly appreciate someone's help clarifying this. 

Thought #1: Spreads exist because of a tenor mismatch. The n-year SOFR swap rate roughly expresses the expected (annualized) compounded overnight repo rate for n years. The n-year treasury, on the other hand, is an n-year loan which carries inflation/duration risk not present in an overnight loan. So the n-year treasury rate should trade higher. 

Thought #2: Suppose an investor wants leveraged long exposure to the rates market and is deciding between: (1) receiving on a SOFR swap and (2) buying treasuries with leverage. Since the floating leg of the swap reflects the overnight financing rate of the treasury, this investor should prefer the leveraged treasury position because of the higher yield (floating legs of both trades are the same so just pick the higher rate). What's the catch here? Something to do with having to refinance the treasury position every day? Are swaps just more convenient for some reason? 

There is also the arbitrage angle (paying on the swap, long the treasury, pocket as much as 70 bps for the 30y but have to hold until maturity). 

Apologies if this question is elementary.