MongoDB, Inc. MDB

Key data

The business

MongoDB sells a database that stores information as flexible documents rather than the rigid rows and columns of the old relational systems. Developers reach for it at the start of a project because the data can change shape as the app does, and once an application's data lives inside it, moving out means rewriting the app. That is the switching cost, and it is real. The money comes two ways: Enterprise Advanced, a license for companies that run the database on their own machines, and Atlas, the managed cloud version that MongoDB runs for you and meters by usage. Atlas is the majority of revenue and the whole story: it grows without a new signature every time a customer's traffic rises. The thing a customer actually holds is a bill that goes up quietly as their app gets busier.

The numbers

The quarters show a company crossing from years of red into black, and doing it while re-accelerating.

Revenue in the July quarter was up 30.5% on the same quarter a year earlier, faster than the 23% the full prior year grew, and profit climbed to $41 million, the largest the company has ever posted. The first profitable quarter arrived in Q4 FY26 at $15.5 million; Q2 FY27 is the second consecutive profitable quarter and nearly triple the size. Non-GAAP earnings have beaten the analyst number four quarters running, the last by a wide margin. This is the inflection: growth stopped decelerating and turned back up in the same stretch that profit arrived.

Revenue compounded about 30% a year from FY2022 to FY2026, and the loss narrowed every single year on the way. Cash tells the cleaner story: operating cash flow went from $150 million in FY2025 to $505 million in FY2026, and the first quarter of this year alone threw off $202 million.

Here is the catch a software owner learns to check first. The 20% free-cash-flow margin is real, but stock handed to employees also runs at 20% of revenue, so the profit that reaches an owner's pocket is far thinner than the cash statement flatters. What the market does not yet price is that the July quarter's re-acceleration is the trend and not a bounce; the single print that settles it is Atlas consumption growth on the next two reports.

Management

Insiders sold and did not buy. Over the last year officers and directors sold $57.5 million across 80 transactions with zero purchases, led by board member Roelof Botha of Sequoia at $15.4 million and CEO Dev Ittycheria at $3.4 million; plan status is not disclosed in the filings I can see, so read the pattern, not any single ticket. Against that, the company bought back stock for the first time ever, $400 million in FY2026 and another $100 million last quarter. But with stock comp at a fifth of revenue, those buybacks mostly mop up dilution rather than shrink the share count, which is the number to watch. The operating record is the redeeming line: four straight beats and a decade-long march from heavy losses to cash generation.

How it fails or surprises you

Consumption cuts both ways. Atlas is metered by usage, so if customers' apps slow, revenue slows the same quarter with no renewal lag to cushion it. Deferred revenue already slipped to $448 million from $471 million. The print that shows it first is sequential Atlas consumption growth on the next call; a soft one arrives instantly, not a year late.

Stock comp eats the owner. The cash margin looks like 20%, but employees take a matching 20% of revenue in stock, so real per-share profit is fragile and one bad quarter can push GAAP earnings back to zero. The buybacks plug the leak rather than drain it. The print: diluted share count flat or falling across a full year, which it has not yet done.

AI-native demand (right tail). If the flood of AI-built applications defaults to MongoDB's flexible model for storing messy, unstructured data and agent memory, net-new workloads land on Atlas and the bill compounds on its own. The market pays for roughly 23% growth today, not 35%. The print: workload adds and revenue growth holding above 30% for consecutive quarters.

Closing thoughts

A specific print settles this one, so name it. The argument is entirely whether the July quarter's jump back to 30% is the new trend or a one-off, and the next two quarters of revenue and Atlas consumption growth answer it cleanly; an ambiguous answer looks like growth drifting into the mid-20s, at which point the reader waits for the third quarter rather than guessing. Survival is not in question: $1.0 billion of cash, almost no debt, and 20% cash margins mean this company decides its own pace. The downside is dead money, growth fading to the high teens and the multiple stuck at 10 times sales while stock comp keeps real profit slim. The upside, which I judge the fatter tail given the re-acceleration already in hand, is a re-rating on both faster growth and a higher multiple from a starting point near the cheapest this stock has ever been. Call that a lean, not a certainty.

The bet is still that companies keep building new software on MongoDB and running more of it on Atlas, with the bill climbing as their apps get busier. It breaks if consumption growth rolls back to the high teens and stock comp keeps GAAP profit hovering near zero. The pair that tells you first, quarter by quarter: Atlas consumption growth and the diluted share count.