Seems pretty reasonable to me. 
Even if you set aside all the graph functionality, Dgraph gives you scalability that would otherwise cost you thousands in dev hours to replicate. If $40/month is a lot then you’re not a “big app”. Most products I’ve worked on with moderate traffic and server load are paying at least $500/month for backend subscriptions (servers, databases, billing providers etc), $40 is a minor footnote in the accounting for even a moderately successful web app.
It’d be nice if there was a $10 tier that was limited to 200mb or something, however, Dgraph have also said that they are under-resourced right now, and they’ve only recently started generating revenue. If they’re under-resourced it makes a lot of sense to focus on the customers who are willing to pay the most money, or at least filter out low-paying customers—because their customer support to revenue ratio is gonna be more favourable to their situation. Dgraph raising their prices is gonna let them move faster and succeed, which is more important to my business in the long term than saving a few hundred dollars in the short-term.