Segwit failed because most bitmain miner with ASICBOOST patented advantage don't want to give up such advantage.

This is simply not true; what was needed in order to activate SegWit was always a promise of a block size increase. This was clear already from the roundtable agreement on 2016-02-21. In general, the miners see very little benefit from SegWit, but they do understand that the block size needs to be increased.

If ASICBOOST was the reason why 75% of the miners rejected SegWit at some point, then mining is already far too centralized, and Bitcoin is fucked up beyond any reason, come SegWit or not.

The point of cryptocurrency is not just about having fast transaction, any company can do that, but most importantly DECENTRALISATION, which is a matter as the main BLOCKCHAIN grows exponentially with blocksize, and if the rate of blockchain size is not taken care of carefully, in long term, normal users wouldn't be able to host the node as the blockchain size growth rate excceed their capable level, that supports decentralisation, thus making the whole blockchain in the hand of big player, which does not help decentralisation at all

The narrative that "bigger blocks causes centralization" is also simply not true. The current capacity restriction causes centralization, stops people from storing Bitcoins in their own wallets, and is generally strangling Bitcoin as a decentralized peer-to-peer network.

It is probably true that, left completely unchecked, the block chain would grow exponentially. It's not at all true that the blockchain grows exponentially with blocksize. It's not important that ordinary users run their own nodes. Most ordinary users (certainly including me) has probably already paid enough in transaction fees to cover hardware and bandwidth costs for running a full node with 8 MB blocks for several years.