Good question.

Let's consider two scenarios:

**Scenario 1**
There is no hard fork. Exchanges and Steemit Inc decide to unvote Tron witnesses (Binance already did it). I suppose that there will be some type of agreement, and previous witnesses will come back.
This is showing me that there is still a power that can take control again in the future. People will not like this, and they will want to sell.

**Scenario 2**
There is a hard fork.
- Assume that we remove the Steem Power of Steemit Inc. Then people will feel that the exchanges could take the control again in the network. They will want to sell.
- Then assume that we remove the tokens from these exchanges. A lot of investors will lose their money. Remember that the price of steem is thanks to investors, then we need a plan for them.
- Ok, we need a plan for these investors, then exchanges still have tokens but we restrict the power up. In this case the exchanges could distribute these tokens between several accounts, difficult to trace, and use these new accounts to power up and take again control. We need a different plan.
- Ok, lets go with this plan: All tokens in these exchanges will lose the right to vote for witnesses. If an investor sends this token to another one then the receptor will lose the right to vote for witnesses using this token. I mean, we create 2 accounting systems: One that uses normal Steem, and another one with Steem LF (limited features) that have lose the right to vote for witnesses. I already proposed this to some witnesses, but there is no good feedback from them, principally because we are creating a new token, which will confuse the people and it's not good for mass adoption.

So, in conclusion it is not clear how to make the hardfork. Then most probably the first scenario will survive.