Selling to large entities, even SaaS, still involves negotiation. Often the large customer wants some key features; sometimes it's a checkbox feature, sometimes it's to disarm antibodies inside the organization, sometimes it's something they actually need for a business requirement. That work needs to be priced in.
And there's the other side of it: different organizations get different amounts of value from our software. If we don't charge organizations who get more value more money, we're leaving profit on the table. Even shrinkwrapped software back in the day had segmented offerings to capture more value. And when you're selling to large organizations, which will necessarily involve multiple salespeople and a bunch of meetings (i.e. scope for fine-grained segmentation without confusing proliferation of products), if you don't charge different prices to different organizations, you're an idiot.
If anything, this way of selling is better for smaller organizations than the segmented approach. Even if we can't charge more, all customers at least get the same product.
That's not to say that volume and data throughput doesn't form a part of the price calculation. But it's a multiplier where the base price is based on the value delivered, which is business dependent.
That's exactly why SaaS disrupt traditional enterprise on-premises software - too complex (installation, user interface), too expensive upfront.
If the software requires no 3 weeks installation phase, no handbook nor training and costs a fraction of the usual 100k to 1m, like $200/host/month, then there is often a clear winner.