Is there a reason big Bitcoin companies wouldn't use only SegWit transactions?

Question

As I understand, SegWit allows anyone to create a SegWit address where they can send their bitcoins and transact from that wallet just as easily as they would with a traditional address. The SegWit inputs, however, would have a lower "weight", therefore they would appear smaller to the miners and thus would need lower fees to be included in a block. This would be despite the transaction itself being a bit bigger byte-wise due to the extra data required for the Witness part of the transaction.

So given all of that, is there any reason big Bitcoin companies that create a lot of transactions on daily basis wouldn't switch to SegWit addresses for the majority of their operations? It looks to me that if they would move their hot wallets onto SegWit addresses they could potentially cut down their transaction fees by a noticeable margin.

Is there any reason this isn't a standard way of doing things yet?

Answer 1

I would like to give one of the reasons (Disclaimer: I am associated with Blockonomics).

There is no support of segwit change addresses yet in bitcoin core. There are workarounds that need to be done to achieve this; as was done by Yodice in this thread. Also to create segwit receiving address one needs to use addwitnessaddress. Lot of businesses just directly use the core wallet without doing lot of programming (there is fear of what may go wrong).

What is needed is that bitcoin core wallet works by default segwit without lot of tinkering. I hear 0.15.1 is better. Looking forward to that

Answer 2

(Disclaimer: I work at BitGo.)

We rolled out segwit support on September 13th for all of our customers. To start using segwit, customers had to only perform a software update and in some cases update the API call when they were previously explicitly requesting non-segwit addresses.

By September 19th, our customers were creating more than 13% segwit transactions. Early in December it was about 63%.

My experience talking to customers was that they frequently were unaware how much savings they would have by switching over. Several times, my presenting them an accurate estimate caused them to switch over within days. A few customers are knowingly lagging behind, because they didn't want to update to a new version or had other reasons why they couldn't use segwit.

Besides, many companies sit on high counts of non-segwit UTXOs, which they'd only cycle through over time.

What I'm trying to say is, even when it's made dead simple to switch over, adoption can take a while. When however a company has to make the transition by themselves it can be quite intimidating, and I could see why it is put off by some that run their own software.