If they do actually beat microwave towers, HFTs will be their first and most enthusiastic customers. SpaceX would not even need any specific strategy for the money to come from HFT. For biggest revenues, add steeply tiered pricing for the best latency, maybe even a bidding process.

If they tried to keep those gains in-house and did their own HFT I would not be surprised if exchanges then found ways to define rules that weaken HFT as a whole. Just a gut feeling that you need to be part of the culture surrounding the exchanges to be in that game and keep winning.

Hybernia from NYC to London is about 55ms whereas I remember reading somewhere these satellites would have a 30ms ping. So probably not for cross-Atlantic propagation once you add all the switching and hops, but maybe cross-Pacific might be doable since propagation delay in space is lower than in fiber.

Realistically, no. High frequency trading is a niche and not hugely profitable market. Not really worth disrupting when there is a much bigger opportunity in delivering internet access to billions of people and many billions of devices.

Interesting point. However, I believe frequency trading relies on connections with extremely low latency, while satellite links are characterized by high Round Trip Time.