The only companies I’ve seen be successful starting with the enterprise have older founders who have spent a decade or more working in the industry and already have the professional network and clout to reach their buyers. Without that professional network and a track record of delivering, you’re not going to make it past the front door.
And in these types of environments, consulting / maintenance absolutely must be part of the business model. Your customers will demand it (or at least demand you have SIs lined up as implementation partners).
Often the problems 'enterprise software' solves aren't overly difficult, and if every company worked the same way it would be trivial to write the software, but the issue is that every company is different, and building everything custom isn't scalable or supportable.
The issue with this is that it means your software won’t meet that clients requirements.
If you want to go into the enterprise space early rather than pivot their later, you can’t start by point blank refusing requirements from your first big client - particularly as you don’t know what customers want yet (or more accurately - your customers are telling you but you are saying it doesn’t meet your vision).
The problem is that if you want to break into enterprise and aren’t willing to make things bespoke, it probably means that your MVP is going to be huge for certain types of software.
This isn’t a rule for all types of software and things like infrastructure provisioning / internal IT tools are different - but I can see where he is coming from.
It’s also different in terms of selling to the SMB or consumer market, where the requirements will either be less-varied (consumer) or the customer will generally sacrifice a lower match to their functional requirements to save cost (SMB). This of course depends on how mature the market is, and in a mature market the MVP gets bigger (for example, word processors).