Because the speculative version is much less trustworthy - and because this kind of service is prone to data lock-in, trust in pricing/profit-models is particularly important.
Consider this: for an existing service with lots of data to back up its pricing, a price list can be treated as an effective guarantee of what the service will continue to cost into the long-term. The company knows that the pricing model is profitable - and therefore, raising prices won't be worth the massive customer outcry.
For a new company, a price list is speculative. There is no way to do accurate long-term extrapolation of profitability. If the prices turn out to be too low to sustain the business, they will have to be raised - and while there will be an outcry, the business will simply have no alternative.
And since it's really hard to migrate all your data/workflows/integrations/etc. away from the service you first decide on, you're incentivised to never join a new service - because the price instability might come back to bite you, let alone if the company itself doesn't collapse later.