MongoDB Q2 FY2027: operating leverage, as SBC grew 6% and revenue 30%

MongoDB sells a database — the general-purpose document data platform that developers reach for instead of a traditional relational database, delivered mostly as a managed cloud service called Atlas. In the quarter ended 31 July it sold $771.8 million of it, up 30% year over year, which management called the fastest growth in several years, and beat consensus on both the top and bottom line.

The beat is the least interesting thing in the release. For years the knock on high-growth software has been that the “profit” in the headline is a non-GAAP number built by adding back an enormous pile of stock-based compensation — the pattern you see at CrowdStrike, where the add-back was larger than the entire non-GAAP profit, and at Palantir, where the same debate runs. MongoDB has that gap too: its headline $1.90 of non-GAAP EPS is 3.8 times the $0.50 it earned under GAAP. What changed this quarter is why the gap is moving. Non-GAAP operating margin went from 15% to 24% in a year, and it did so because stock compensation grew 6% while revenue grew 30%. That is operating leverage, and it is the one thing in this print worth understanding.

The results

	Q2 FY2027	Consensus	Q2 FY2026
Total revenue	$771.8M	~$733.6M	$591.4M
Non-GAAP EPS (fully diluted)	$1.90	~$1.60	$1.00
GAAP diluted EPS	$0.50	—	$(0.58)
Non-GAAP operating margin	24%	—	15%
GAAP gross margin	74%	—	71%

Revenue grew 30.5% year over year. Subscription revenue of $747.1 million (up 31%) is essentially the whole company; services of $24.6 million (up 29%) is the rest. Non-GAAP EPS of $1.90 was up 90% on the year and cleared the roughly $1.60 the Street carried by about thirty cents; revenue beat the ~$733.6 million consensus by around 5%. Against its own prior-year figures this is a beat on every line that matters, and management raised the full-year outlook on the back of it.

Did MongoDB beat its own guidance in Q2 FY2027?

Yes, on every guided line. Revenue of $771.8 million came in above the top of the $729–$734 million range MongoDB guided to in its May 28, 2026 8-K, and non-GAAP EPS of $1.90 was above the top of its $1.58–$1.61 range. Non-GAAP operating income of $185.9 million cleared a $152.0–$156.0 million guide, and GAAP operating income of $28.4 million compared with a guided loss of $(10.0) million to $(6.0) million. Year over year, revenue rose 30%, and the full-year revenue outlook went to $2.99–$3.03 billion from $2.92–$2.96 billion in May.

Where the revenue comes from

Product line	Q2 FY2027	Share	YoY
Atlas-related	$565.9M	73%	+29%
Enterprise Advanced & other	$181.2M	24%	+36%
Services	$24.6M	3%	+29%
Total	$771.8M	100%	+30%

Atlas — the managed cloud database — is 73% of revenue and remains the center of gravity. But the line that grew fastest this quarter was not Atlas. Enterprise Advanced (EA) and other, the self-managed software customers run in their own or private-cloud environments, grew 36%, ahead of Atlas’s 29%. That matters because the long-standing bear worry on MongoDB was that Atlas would cannibalise EA — that the cloud product would eat the on-prem product and leave the company with one growth engine, not two. This quarter both grew, and the slower-shrinking one actually reaccelerated. Management’s framing — a “run anywhere” strategy spanning cloud and self-managed — is, for this quarter at least, visible in the numbers rather than just the slide.

The customer base backs the growth up rather than flattering it: total customers reached 70,600+, up from 59,900+ a year ago (about +18%), and customers spending $100,000 or more of annualised recurring revenue rose to 2,999 from 2,564 (about +17%). Growth is coming from more customers and larger ones, not a handful of outsized deals.

The one number: leverage, not the beat

Here is the same business, GAAP and non-GAAP, this year against last.

	Q2 FY2027	Q2 FY2026
GAAP income (loss) from operations	$28.4M	$(65.3)M
GAAP operating margin	4%	(11)%
Non-GAAP income from operations	$185.9M	$86.8M
Non-GAAP operating margin	24%	15%
Stock-based compensation	$148.9M	$140.4M
SBC as % of revenue	19%	24%

Read the last two rows first, because they are the story. A year ago MongoDB spent 24% of revenue on stock compensation. This quarter it spent 19% — not because the dollar figure fell (it rose slightly, to $148.9 million) but because revenue grew 30% while SBC grew only 6%. When your single largest non-cash cost grows at a fifth of the pace of your revenue, margins expand almost mechanically, and they did: non-GAAP operating margin went from 15% to 24% in twelve months, and GAAP operating income flipped from a $65.3 million loss to a $28.4 million profit.

This is precisely the thing the SBC-heavy software model is supposed to eventually produce and often doesn’t. At Marvell last quarter, stock compensation grew 112% — three times revenue growth — so the dilution was widening, not narrowing. At CrowdStrike it ran at 27% of revenue. MongoDB at 19% and falling, with SBC growing 6%, is the counter-example: the add-back is shrinking as a share of the business. The gap between GAAP and non-GAAP is still there, but for once it is closing from the right end.

The GAAP-to-non-GAAP bridge, and an honest note on tax

The gap is real and worth walking, because $0.50 and $1.90 are very different numbers.

GAAP → non-GAAP net income	Q2 FY2027
GAAP net income	$40.9M
+ Stock-based compensation (incl. payroll taxes)	$153.4M
+ Amortisation of intangible assets	$3.2M
+ Legal fees (securities litigation)	$0.6M
+ Acquisition-related and other	$0.3M
− Gains on financial instruments	$(0.2)M
− Income-tax effects and adjustments	$(35.6)M
Non-GAAP net income	$162.6M

Two things are worth flagging. First, the dominant add-back is stock compensation — $153.4 million (the income-statement SBC of $148.9 million plus employer payroll taxes on vesting), and without it there is no non-GAAP story. That is the standard shape and the standard caveat: it is a real cost paid in shares, it dilutes owners, and the headline EPS is built to exclude it.

Second, and to MongoDB’s credit, the bridge subtracts $35.6 million of tax. MongoDB’s non-GAAP net income applies an assumed long-term 20% tax rate, while its GAAP tax bill this quarter was only $5.0 million. So the non-GAAP number is taxed more heavily than the GAAP number — the opposite of the aggressive move, where a company adds back costs and leaves the tax benefit untouched to flatter the result. It is a small point, but it is the difference between a non-GAAP number engineered to look as large as possible and one that at least normalises the tax it excludes.

The dilution itself is being managed rather than ignored. Diluted GAAP shares rose only about 1%, from 81.1 million to 82.0 million, and the non-GAAP fully diluted count actually fell, from 87.1 million to 85.8 million. MongoDB net-settled equity awards — paying $59.4 million of employees’ tax withholding in cash rather than issuing the shares — and repurchased $100.0 million of stock in the quarter ($200.3 million for the half). The balance sheet shows the effect: additional paid-in capital and treasury stock both moved as shares were retired. Buying back stock to offset comp is cash out of the door to stand still, but combined with the slow SBC growth it is why the share count is roughly flat instead of climbing.

Cash, backlog, and the balance sheet

The cash flow is the cleanest confirmation that the leverage is real and not an accounting artefact.

	Q2 FY2027	Q2 FY2026
Operating cash flow	$141.9M	$72.1M
Free cash flow	$137.6M	$69.9M
FCF margin	17.8%	11.8%

Free cash flow nearly doubled year over year, and the FCF margin expanded from 11.8% to 17.8% — the same leverage the margin lines show, but in cash, where there is no add-back to argue about. MongoDB ended the quarter with $2.4 billion in cash, cash equivalents and short-term investments and no long-term debt on the balance sheet; the convertible notes that once sat there are gone. This is a company funding its own buyback and growth out of cash generation, not a balance sheet.

One forward indicator stands out: remaining performance obligations rose 91% to $1.52 billion, and the current portion (cRPO, the piece expected to convert within a year) rose 73% to $797.3 million. RPO is contracted-but-unrecognised revenue — backlog — and a 91% jump against 30% revenue growth points to larger, longer, committed deals, consistent with the EA reacceleration. It is not revenue yet, and multi-year contracts can lump RPO around, but it is the least ambiguous sign in the release that the pipeline is deepening.

The bear case, stated fairly

The quarter was strong; the honest reservations are narrower than “it’s overvalued.”

1. The Q3 revenue guide is a sequential step down. Management guided the third quarter to $756–$761 million, a midpoint below this quarter’s $771.8 million. MongoDB has a long history of guiding Atlas consumption conservatively and beating, and it raised the full-year outlook to $2.99–$3.03 billion — so this is very likely sandbagging, not a warning. But it is what the company put on paper, and the reader should know the guide implies no sequential growth next quarter.
2. Full-year GAAP operating income is still guided to a loss. For all the leverage, the FY2027 GAAP operating guide is $(28) million to $(8) million — a loss at both ends. The GAAP net profit ($0.53–$0.77 for the year) leans on other income, chiefly interest on the $2.4 billion cash pile. Operating profitability on a GAAP basis is real this quarter ($28.4 million) but not yet durable across the year on the company’s own numbers.
3. The non-GAAP EPS still excludes $150 million a quarter of a real cost. Leverage improving is not the same as the cost disappearing. SBC is still 19% of revenue, and the $1.90 headline is 3.8 times the GAAP $0.50. The trajectory is good; the absolute gap is still large.
4. AI is a thesis, not yet a line item. Much of the release is about AI retrieval, vector search, and an MCP server connecting coding agents to Atlas data. These are credible product moves, but the 30% growth this quarter came from “core enterprise workloads,” in management’s own words — AI is described as “early momentum,” not a quantified driver. The AI story is a reason to own the durability; it is not yet in the numbers.

The bull case

Also on the page, and concrete. Revenue grew 30%, the fastest in years, and beat by 5%. Both product lines grew, with the self-managed one reaccelerating to 36% — answering the cannibalisation worry directly. Non-GAAP operating margin expanded nine points to 24% while SBC grew just 6%, and free cash flow nearly doubled to a 17.8% margin. RPO jumped 91%, the balance sheet is debt-free with $2.4 billion of cash, the share count is roughly flat, and management raised the full-year revenue outlook. This is a company demonstrating, in cash and in GAAP operating income, the profitability that its category is usually only promised.

What to check at the Q3 report

• Revenue against the $756–$761 million guide — specifically whether the sequential-flat guide was conservatism (a beat) or a genuine Atlas-consumption slowdown.
• Atlas growth against 29%, and EA & other against 36% — whether the self-managed reacceleration holds or reverts, and whether Atlas stays the larger base.
• SBC as a percent of revenue against 19%, and its growth rate against 6% — the single dial behind the margin story. If SBC growth stays in the single digits, the leverage compounds.
• Non-GAAP operating margin against 24% — whether the nine-point year-over-year expansion continues or plateaus.
• Free cash flow margin against 17.8% — the check that the leverage is cash, not accounting.
• RPO and cRPO against $1.52 billion / $797.3 million — whether the 91% backlog jump was a one-quarter multi-year-deal spike or a new run-rate.

The read

MongoDB delivered a genuinely strong quarter — 30% growth, a beat on both lines, both product lines expanding, and a raised full-year outlook — but the beat is not the point. The point is that this is what operating leverage looks like when it finally arrives in a stock-comp-heavy software business: SBC growing 6% against 30% revenue growth, non-GAAP operating margin from 15% to 24%, GAAP operating income turning positive, and free cash flow nearly doubling. The $1.90-versus-$0.50 gap between the two EPS figures is still wide, and stock compensation is still 19% of revenue — nobody should pretend the add-back stopped mattering. But it is shrinking as a share of the business, and it is shrinking in cash as well as in non-GAAP margin, which is the version that can’t be argued with.

For anyone holding MDB, the two numbers to watch next quarter are stock-based compensation as a percent of revenue — does the 19%-and-falling trend continue, because that is the entire leverage story — and revenue against the flat Q3 guide, because that tells you whether the 30% growth was a peak or a floor. Growth tells you the platform is wanted. The SBC trend tells you how much of that growth reaches the owner.

---

Sources and method. Company figures in this piece — total revenue of $771.8 million against $591.4 million a year ago, subscription revenue of $747.1 million and services of $24.6 million; the product split of Atlas-related revenue of $565.9 million, Enterprise Advanced and other of $181.2 million and services of $24.6 million; GAAP gross profit of $569.8 million (74% margin) and non-GAAP gross profit of $585.7 million (76%); GAAP income from operations of $28.4 million against a $65.3 million loss, and non-GAAP income from operations of $185.9 million (24% margin) against $86.8 million (15%); GAAP net income of $40.9 million ($0.51 basic, $0.50 diluted) against a $47.0 million loss, and non-GAAP net income of $162.6 million ($1.90 fully diluted) against $87.2 million ($1.00); total stock-based compensation expense of $148.9 million (and the $153.4 million add-back including employer payroll taxes) against $140.4 million; the income-tax-effects adjustment of $35.6 million and the assumed 20% long-term non-GAAP tax rate; operating cash flow of $141.9 million and free cash flow of $137.6 million against $69.9 million; $2.4 billion of cash, equivalents and short-term investments; remaining performance obligations of $1,519.2 million (+91%) and cRPO of $797.3 million (+73%); customer counts of 70,600+ total, 69,300+ Atlas and 2,999 customers over $100K ARR; $100.0 million of share repurchases in the quarter ($200.3 million for the half), $59.4 million of taxes paid on net share settlement, weighted diluted shares of 82.0 million GAAP and 85.8 million non-GAAP fully diluted; and all Q3 and full-year FY2027 guidance (Q3 revenue $756–$761 million; FY revenue $2.99–$3.03 billion; the GAAP operating loss guides and the non-GAAP EPS ranges) — are MongoDB’s own disclosures from its Q2 FY2027 results release (Form 8-K, Exhibit 99.1) filed 2026-09-01 for the quarter ended 2026-07-31, reported after the close. Prior-year comparatives are the year-ago columns in that same release, and CJ Desai (President and CEO) and Mike Berry (CFO) are quoted from it.

Not company figures, and attributed accordingly: the consensus estimates (revenue ~$733.6 million and non-GAAP EPS ~$1.60, from the Zacks consensus reported in the previews cited above). Separately from those outlet figures, the guidance verdict is judged only against MongoDB’s own previously issued guidance (Exhibit 99.1 to the Form 8-K filed 2026-05-28, accession 0001628280-26-038798, which also carries the prior full-year outlook) and against the year-ago quarter. This piece makes no claim about the share-price reaction.

Our arithmetic, labelled as such: the 30.5% total and 30.5% subscription revenue growth rates; the product-line YoY rates (Atlas +29%, EA & other +36%) and their shares of revenue (Atlas 73%); SBC at 19% of revenue this year versus 24% a year ago and its 6% growth rate against 30% revenue growth; the 3.8× ratio of non-GAAP to GAAP EPS; the 17.8% versus 11.8% free-cash-flow margins and the “nearly doubled” cash-flow growth; the ~18% and ~17% customer-count growth rates; the roughly flat diluted share count; and the sequential read on the Q3 revenue guide. These are approximations built on the disclosed figures, not reported numbers. The interpretation — that the story is operating leverage rather than the beat, that MongoDB’s shrinking SBC-to-revenue ratio is the counter-example to the widening one at other software names, and that the tax-adjusted non-GAAP bridge is less aggressive than an add-back-only approach — is our analysis, not a company statement, and is not a forecast.

Related: CrowdStrike Q2 FY2027, where the same GAAP-versus-non-GAAP gap runs the other way — a stock-comp add-back larger than the entire non-GAAP profit; Palantir Q2 2026, the other high-growth software name where the profit-lens debate is the whole story; Microsoft Q4 FY2026, whose Azure cloud is both MongoDB Atlas’s largest host and its largest competitor; and Shopify Q2 2026, another platform business where operating leverage and free cash flow are the numbers that settle the argument.

Tradune shows what you actually hold across every account, and whether a position has beaten a plain index fund since you bought it. Educational analysis, not investment advice.