MongoDB, Inc.
MongoDB sells a document database and Atlas, the version that runs in its cloud and is billed mostly on use. PostgreSQL, Oracle and the document databases sold by Amazon, Microsoft and Google do similar work. It is not the most popular database by decree. An artificial-intelligence model can be trained on data that lives somewhere else. Atlas is not that training.
An application already written sticks, and it can be rewritten. The customer can leave Atlas and keep running the database on its own servers. This report does not count programmers or a migration rate. If the customer spends less on cloud, they usually spend less on Atlas. A year of heavy development is not normal earnings.
Cash is the subscription and that use, after the cost of the cloud. It has to fit. It is not a toll on every record the company keeps.
"The advantage is the application already in production on Mongo and the team that already runs it on Atlas. The customer can migrate or install the database elsewhere. The moat narrows if usage falls, or if another database holds the same document."
Loading the Vaultflake…
—
Health is debt against a usage-based software business, not against the customer's servers. Quality separates the subscription from Atlas consumption, after the cost of that cloud. Growth follows usage and new accounts, not a licence over the data. P/E is read against a year of heavy development, not against that peak. Shareholder return, if any, comes out of that cash. Do not read it as millions of programmers already locked in.
The P/E of a year of heavy Atlas use is not normal earnings. EV/EBITDA is read without treating the customer's consumption as assured revenue. Do not use a cloud's multiple, which sells the server, or another database's multiple as if the book were the same. A dash if the MDB cache is cold.
Not advice on MongoDB. Vaultflake does not count programmers or treat Atlas as the place a model has to live. Read the filings. The table is a snapshot, not a target price.
The advantage is the application already in production on Mongo and the team that already runs it on Atlas. The customer can migrate or install the database elsewhere. The moat narrows if usage falls, or if another database holds the same document.
Health is debt against a usage-based software business, not against the customer's servers. Quality separates the subscription from Atlas consumption, after the cost of that cloud. Growth follows usage and new accounts, not a licence over the data. P/E is read against a year of heavy development, not against that peak. Shareholder return, if any, comes out of that cash. Do not read it as millions of programmers already locked in.
The P/E of a year of heavy Atlas use is not normal earnings. EV/EBITDA is read without treating the customer's consumption as assured revenue. Do not use a cloud's multiple, which sells the server, or another database's multiple as if the book were the same. A dash if the MDB cache is cold.
Not advice on MongoDB. Vaultflake does not count programmers or treat Atlas as the place a model has to live. Read the filings. The table is a snapshot, not a target price.
Technology · Consumer Devices
Technology · Software and Cloud
Technology · Semiconductors
Technology · Semiconductor Equipment
Technology · Business Software
Technology · Database and Cloud
A free account opens the interactive chart. The Vault assistant is Premium.