MongoDB is back. It saw growth slow materially for a while, to as low as 13% … but it’s back. And it’s now an AI Beneficiary. At $2.4 Billion in ARR, growth has reaccelerated to 24%.
And the stock is up a stunning 45% this year.
5 Interesting Learnings:
#1: 70% of Atlas ARR Comes from Customers Using Multiple Platform Capabilities Beyond “Database
– And Those Customers Have 5x Higher ARR
MongoDB disclosed that 70% of their total Atlas ARR now comes from customers using at least one additional platform capability beyond the core database. Not 30%. Not 50%. 70%.
But here’s the part that matters even more: customers who use at least one additional capability have 5x higher Atlas ARR compared to those who don’t.
Think about what this means in practical terms. If you’re a customer spending $100K annually on just MongoDB’s core database, and you add Vector Search or Time Series or any other capability, MongoDB’s data suggests you’re statistically more likely to be spending $500K.
This isn’t just standard NRR expansion. This is platform economics at work. The customers who see MongoDB as a platform rather than a point solution are fundamentally different customers with fundamentally different budget envelopes.
What’s driving this? MongoDB’s bet is that once you’re building on their document model and you need adjacent capabilities, the switching cost and integration complexity of using separate vendors becomes prohibitive. So you consolidate onto the MongoDB platform.
At $2B+ in revenue, having 70% of your Atlas business coming from multi-product customers isn’t a science experiment anymore. It’s proof of concept at massive scale.
#2: The FY22 Cohort of >$100K Customers Has Expanded to 200% of Original ARR After Just 3 Years
Cohort expansion data rarely gets disclosed this explicitly, especially for customers already past $100K.
MongoDB shared that customers who crossed $100K ARR in FY22 have expanded to an average of 200% of their original ARR by FY25. That’s 2x in three years. Not total customer base expansion – this is same-cohort expansion after they were already spending six figures.
For context, the >$1M ARR cohort from FY22 expanded to 185% over the same period. Still impressive, but slightly lower. This suggests the $100K-$1M segment might actually have better expansion dynamics than the $1M+ segment, which is counterintuitive to how most people think about enterprise expansion.
The implication here is profound: once a customer commits to spending $100K+ with you, you have a three-year window where they’re likely to double. That fundamentally changes how you should think about CAC payback and customer lifetime value at scale.
If your $100K customer typically costs you $30K to acquire and expand, and they double in three years, you’re looking at $200K ARR with $30K CAC by year three. That’s a 6.7x return in 36 months before factoring in any expansion beyond year three. Those are venture-scale returns on sales and marketing spend.
But here’s the question this raises: what happens in year four and beyond? Does expansion slow down? Plateau? Accelerate? MongoDB didn’t share that data, which means we should assume it’s either not as compelling or still too early to have meaningful data on the FY22 cohort beyond year three.
#3: 30% of Atlas ARR Now Comes from Customers with at Least One AI Use Case
MongoDB disclosed that ~30% of Atlas ARR now comes from customers running at least one AI use case on their platform. Not total company revenue – specifically Atlas ARR.
Given that Atlas is the vast majority of MongoDB’s revenue at this point (and growing significantly faster than their on-prem business), this means roughly 30% of their high-growth engine is now directly tied to AI workloads.
Let’s do the math. MongoDB’s guiding to approximately $2.3B in total revenue for FY26. If we assume Atlas is roughly 75% of that (based on their historical disclosures), that’s ~$1.7B in Atlas revenue. 30% of that is ~$510M in Atlas ARR directly attributable to customers with AI use cases.
That’s a half-billion-dollar AI business that didn’t exist in any meaningful way two years ago.
What’s particularly notable is the speed. MongoDB has been talking about AI opportunities for less than two years in a serious way. To get to 30% penetration of Atlas ARR in that timeframe suggests this isn’t a slow-burn adoption curve – it’s happening fast.
The question is sustainability. Are these customers building production AI applications that will scale and persist? Or are these still experimental workloads that could evaporate if AI hype cools? MongoDB’s betting on the former, obviously. But 30% exposure to what is still a relatively nascent application category is meaningful risk if the market shifts.
#4: They added 2,500 net new Atlas customers in Q2 FY26 – the second-highest quarter in the last six years
Customer acquisition isn’t slowing down even at scale, and the timing makes this particularly interesting.
MongoDB added 2,500 net new Atlas customers in Q2 FY26 (the July-October period). This wasn’t just a strong quarter – it was the second-highest customer acquisition quarter in the last six years.
The fact that this is still accelerating (second-best quarter in six years) rather than plateauing suggests MongoDB hasn’t hit any natural ceiling on self-serve customer acquisition yet. That’s rare at their scale and revenue.
#5: 25% of Customers with >$1M ARR Originally Started in Self-Serve – And They Hit $1M 15% Faster
MongoDB’s product-led growth motion is producing genuine enterprise customers at scale.
25% of customers now spending >$1M annually with MongoDB originally came through the self-serve channel. Not direct sales. They started by signing up for Atlas on their own, likely with a credit card, and expanded into seven-figure accounts.
But here’s the more interesting part: these self-serve-originated customers reached $1M in ARR 15% faster than customers sourced through traditional direct sales channels.
This flips the conventional wisdom on its head. Most people assume enterprise customers sourced through self-serve take longer to ramp because they’re starting smaller and have less initial commitment. MongoDB’s data suggests the opposite – at least for the cohort that successfully expands to $1M+.
Why might this be true? A few theories:
First, self-serve customers are pre-qualified by definition. They had a problem, found MongoDB, and adopted it without sales assistance. That suggests technical fit and user motivation are already validated before any sales rep gets involved.
Second, self-serve customers often start with production workloads immediately. They’re not going through a six-month enterprise sales cycle – they’re solving a problem today. So by the time sales engages with them for expansion, they’re already running mission-critical workloads on MongoDB, which accelerates expansion conversations.
Third, bottoms-up adoption creates organizational buy-in that top-down sales struggles to replicate. If MongoDB is already running in production across multiple teams before the enterprise sales cycle begins, the organization has self-selected for MongoDB as a standard, making expansion easier and faster.
The 15% faster time-to-$1M metric is the part that should make every enterprise SaaS company rethink their go-to-market mix. If product-led growth isn’t just generating leads but actually accelerating enterprise expansion velocity, the ROI math on investing in self-serve infrastructure becomes significantly more compelling.
More Interesting Learnings That Didn’t Make the Top 5
• AMP (Relational Migrator) achieved 50x faster code conversion and 90% accuracy on migrations – MongoDB’s tooling for migrating from relational databases to MongoDB is showing real traction. The 50x speed improvement suggests they’re using AI/LLM-based code conversion, which makes sense. The 90% accuracy is the more interesting number – it means 10% of migrations still require manual intervention. That’s good enough to massively accelerate migrations but not good enough to be fully automated, which creates services revenue opportunities.
• Operating margin expanded from 12% to 14% YoY while maintaining 29% Atlas growth – MongoDB is proving you can drive margin expansion without sacrificing growth at $2B+ scale. This is the math that makes public market investors happy – you’re growing the top line at 20%+ while expanding margins by 200bps annually. If they can maintain that trajectory to $5B, they’ll be generating 20%+ operating margins at scale while still growing revenue in the double digits. That’s Salesforce-level economics.
The overarching theme here is that MongoDB is proving platform expansion economics work at scale. Not in theory – in practice, with real data, at billions in revenue.
The 70% multi-product attachment, the 5x ARR lift from platform adoption, the 200% cohort expansion – these aren’t startup metrics. These are the metrics of a company that’s figured out how to compound growth through platform expansion rather than just customer acquisition.
Most interesting of all might be what MongoDB is demonstrating about AI timing. Getting to 30% of Atlas ARR from AI use cases in less than two years suggests the market is moving faster than most people modeled. That has implications for every database and data infrastructure company – the AI wave isn’t coming, it’s already here, and it’s moving faster than the previous waves (mobile, cloud) that reshaped infrastructure markets.
The question isn’t whether AI will drive database revenue. MongoDB’s numbers suggest it already is. The question is which vendors are positioned to capture it, and which are still trying to figure out their AI story while the market moves past them.
