Written by TFN Research Desk — covering startups, technology, digital media, and business strategy.
While millions of developers fled rigid SQL schemas in the late 2000s to build MongoDB into a publicly traded company, a growing share of them are now migrating straight back to SQL for the AI era.
MongoDB closed fiscal 2026 with $2.46 billion in total revenue, up 23 percent year over year, and a market capitalization of roughly $28.5 billion as of mid-June 2026 (MongoDB Investor Relations, March 2026; CNBC, June 2026). That outcome traces back to a single frustration three DoubleClick engineers had with relational databases nearly two decades ago: SQL’s rigid schemas could not keep pace with how fast their applications needed to change. The database they built to solve that problem became one of the clearest developer led growth stories in enterprise software. It also now sits inside a genuinely contested moment, because the same developers who abandoned SQL for MongoDB’s flexibility are, in increasing numbers, migrating back toward SQL again.
Startup Strategy • Case Study • Database Infrastructure • Developer Tools
The hook: a frustration that became a category
In the mid-2000s, Dwight Merriman and Eliot Horowitz were running into the limits of relational databases while scaling DoubleClick’s advertising systems to serve hundreds of thousands of ad requests per second (MongoDB, “Our Story”). Rigid table schemas meant every change to how data was structured required a slow, coordinated migration across the entire system, exactly the kind of friction that frustrates fast-moving engineering teams. After Google acquired DoubleClick, Merriman and Horowitz teamed up with serial entrepreneur Kevin P. Ryan to found 10gen in 2007, with the explicit goal of building a database that let the application’s data model evolve as fast as the code did (MongoDB, “Our Story”).
Why this story matters
MongoDB’s history matters because it is the clearest proof that a database, normally the most conservative layer of any stack, can still be disrupted by a developer led growth motion if the incumbent experience is painful enough. The same dynamic that let an open source document database take share from Oracle and MySQL is visible again today in how Indian SaaS teams pick their infrastructure stack, frequently starting with whatever tool offers the fastest path from idea to working prototype, as covered in The Founder Nation’s look at how developers became Vercel’s growth engine. With India having added 5.2 million new developers to GitHub in 2025 and overtaken the United States as the largest base of open source contributors, the database decisions this generation makes will shape the next decade of infrastructure spending (Analytics India Mag, October 2025).
Quick facts
| Category | Detail |
|---|---|
| Founders | Dwight Merriman, Eliot Horowitz, Kevin P. Ryan |
| Founded | 2007, as 10gen; renamed MongoDB Inc. in 2013 |
| Core product | MongoDB document database; MongoDB Atlas managed cloud platform |
| Public listing | Nasdaq: MDB, IPO October 2017 |
| FY2026 revenue | $2.46 billion, up 23% year over year (MongoDB Investor Relations, March 2026) |
| Market capitalization | Approximately $28.5 billion as of June 16, 2026 (CNBC, June 2026) |
| Total customers | Over 65,200 as of January 31, 2026 (MongoDB Investor Relations, March 2026) |
| CEO | Chirantan “CJ” Desai, since November 10, 2025 (MongoDB Investor Relations, November 2025) |
| Industry tags | Database, developer tools, cloud infrastructure |
Background
10gen’s original plan was not even a database, it was a full platform-as-a-service. While building that platform, the founding team could not find any existing database that fit the flexible, horizontally scalable architecture they wanted, so they built their own document-oriented storage layer internally (Medium, “Why MongoDB Outperformed Its Competitors,” February 2025). By early 2009, that internal database, nicknamed MongoDB after “humongous,” had clearly become the more valuable asset than the platform built around it, and the team made the consequential decision to scrap most of the original PaaS code and release MongoDB itself as an open source project.
For Indian engineering teams, the appeal of MongoDB’s pitch in its early years mirrors a pattern still visible today: a flexible, JSON-like document model meant a backend engineer could start storing data without first designing a complete relational schema, an attractive proposition for resource-constrained startup teams iterating quickly on product-market fit.
Timeline
- 2007: 10gen is founded in New York by Dwight Merriman, Eliot Horowitz, and Kevin P. Ryan (MongoDB, “Our Story”).
- 2009: MongoDB is released publicly as an open source document database.
- 2013: 10gen renames itself MongoDB, Inc. to align the corporate identity with its flagship product (TechCrunch, August 2013).
- 2016: MongoDB Atlas, the company’s managed database-as-a-service offering, launches.
- 2017: MongoDB completes its initial public offering on Nasdaq under the ticker MDB in October.
- 2018: MongoDB moves away from the AGPL open source license to its own Server Side Public License, a shift that was not approved by the Open Source Initiative.
- 2025: CJ Desai succeeds Dev Ittycheria as President and CEO on November 10, after Ittycheria’s eleven-year tenure (MongoDB Investor Relations, November 2025).
- 2026: MongoDB closes fiscal year 2026 with $2.46 billion in revenue, with Atlas accounting for 75 percent of total revenue by the first quarter of fiscal 2027 (MongoDB Investor Relations, May 2026).
How it happened
Move 1: solve a real engineering pain, not a theoretical one
MongoDB’s document model was not designed in the abstract, it was designed to fix a specific, lived frustration with relational databases at DoubleClick’s scale. That grounding in a real, painful workflow problem gave the product an unusually credible founding story to tell developers evaluating it for the first time, since the people who built it had personally hit the wall it was designed to remove.
Move 2: give the database away to win the bottom-up funnel
Releasing MongoDB as a free, open source project in 2009 let it spread the same way Linux, Postgres, and other foundational infrastructure had spread before it: through individual engineers downloading it, trying it on a side project, and recommending it to colleagues, with no sales conversation required. This open source distribution built the installed base that MongoDB’s later commercial products would eventually monetize.
Move 3: convert the installed base into recurring cloud revenue with Atlas
For years, MongoDB’s commercial model rested on enterprise support contracts and on-premise licensing for the database it gave away for free, a model with a real ceiling. Launching Atlas in 2016 changed the underlying economics entirely, turning a self-hosted open source tool into a usage-based, recurring cloud subscription. By the first quarter of fiscal 2027, Atlas alone represented 75 percent of MongoDB’s total revenue, more than double its 2025 share of total revenue (SEC filing, MongoDB 10-Q, FY2026).
The strategy behind the success
MongoDB’s developer led growth strategy depended on sequencing its monetization correctly: build trust for free through open source distribution first, then monetize a separate cloud layer once the developer base was large enough to make a managed offering valuable. Eliot Horowitz, the company’s co-founder and longtime chief technology officer, has spoken about how the rise of NoSQL was less a marketing label and more a genuine response to the limitations developers were hitting with rigid relational schemas, and has been candid that mastering open source as a sustainable business, rather than just a technology choice, remains one of the harder problems any infrastructure company has to solve (Medium, “MongoDB co-creator explains why ‘NoSQL’ came to be,” January 2016). [VISUAL: infographic showing the funnel from free, open source MongoDB downloads to paid Atlas subscriptions] That separation between the free technology and the paid infrastructure around it is the same structural pattern this cluster has documented at Vercel, and it remains the clearest throughline across nearly every developer-first infrastructure company that has reached public-company scale.
Business model breakdown
MongoDB now runs primarily as a usage-based cloud subscription business through Atlas, layered on top of a free, source-available core database. Atlas customers are billed either on a self-serve, pay-as-you-go basis for smaller usage or through annual contracts sold by MongoDB’s sales force for larger enterprise deployments (SEC filing, MongoDB 10-Q, FY2026). The remaining revenue comes from Enterprise Advanced, MongoDB’s on-premise and self-managed licensing product, which has become a smaller share of total revenue over time but remains relevant for regulated industries and AI workloads that require data to stay on-premise. The economics mirror the same shape as MongoDB’s original open source funnel: low-friction adoption at the bottom, with usage growth, not a separate sales-driven decision, triggering the move into paid Atlas tiers.
Comparison table
Short answer
MongoDB’s $2.46 billion in fiscal 2026 revenue and 23 percent year-over-year growth make it the clear commercial leader among NoSQL databases, but PostgreSQL now leads on raw developer usage, with 55.6 percent of developers using it in 2025 compared with roughly 24 to 26 percent for MongoDB, a reversal of the trend that originally built MongoDB’s growth.
| Dimension | MongoDB | PostgreSQL | Oracle |
|---|---|---|---|
| Data model | Document (NoSQL) | Relational (SQL), with JSON support | Relational (SQL) |
| 2025 developer usage share | Approximately 24 to 26% (Stack Overflow Developer Survey 2025, via Vonng, July 2025) | 55.6%, the highest of any database (Stack Overflow Developer Survey 2025, via Vonng, July 2025) | Approximately 10.6% (Stack Overflow Developer Survey 2025, via Nucamp, January 2026) |
| Ownership | Public company, Nasdaq: MDB | Open source, no single owning company | Public company, NYSE: ORCL |
| Primary monetization | Atlas usage-based cloud subscription | Monetized indirectly via cloud vendors (AWS RDS, Supabase, Neon) | Enterprise licensing and cloud infrastructure |
| Most-wanted database rank, 2025 | Fifth, down from first in 2017 to 2020 (Vonng, July 2025) | First, for the third consecutive year (Vonng, July 2025) | Not in the top five |
[VISUAL: comparison chart rendering the table above as a shareable image]
What competitors missed
Oracle and the broader relational database incumbents missed how much developer experience mattered as a competitive axis in the 2010s. They optimized for administrators and DBAs rather than the individual developers who increasingly had the authority to choose their own stack, which is exactly the opening MongoDB’s free, flexible, code-first database walked through. The more interesting miss runs in the opposite direction today: for years, the relational database community underestimated how far JSON support inside Postgres could close MongoDB’s flexibility advantage without sacrificing transactions, joins, or schema constraints, a gap that newer serverless Postgres players have aggressively closed. As covered in The Founder Nation’s breakdown of [INTERNAL LINK: GitLab, how a fully remote company went public], owning the developer relationship is a moat that erodes the moment a competitor matches the experience while keeping its original structural advantages intact.
Risks and challenges
- Developer sentiment has reversed sharply: MongoDB ranked as the most-wanted database from 2017 through 2020 but had fallen to fifth place by the 2025 Stack Overflow Developer Survey, while PostgreSQL has held the top spot for three consecutive years (Vonng, July 2025).
- PostgreSQL’s JSONB improvements increasingly replicate MongoDB’s original schema-flexibility pitch while keeping ACID transactions, joins, and constraints, narrowing the gap that justified choosing MongoDB over SQL in the first place.
- Wire-compatible tools such as FerretDB let teams run MongoDB drivers and tooling directly on top of Postgres, a direct technical threat to MongoDB’s lock-in.
- The 2018 move to the Server Side Public License, which is not approved by the Open Source Initiative, created friction with parts of the open source community and prompted AWS to build its own MongoDB-compatible DocumentDB as a workaround.
- The company remains GAAP unprofitable, reporting a net loss for fiscal 2026 even as revenue grew 23 percent year over year.
- Revenue is now heavily concentrated in Atlas, making MongoDB’s growth more sensitive to broader cloud infrastructure spending cycles and hyperscaler competition than it was in MongoDB’s earlier, license-driven era.
What founders can learn
MongoDB’s twenty-year arc offers founders a far more nuanced lesson than the simple “give it away free” story usually told about open source companies. The first lesson is that the original commercial model rarely survives long enough to build a public company on its own. MongoDB’s first business, enterprise support and licensing wrapped around a free database, was a real but limited business; the moment that actually transformed the company’s growth trajectory was the 2016 launch of Atlas, a usage-based cloud product that monetized the same installed base in a completely different way. Founders who only validate their first monetization model and never revisit it are leaving the more durable version of their business on the table.
The second lesson concerns differentiation decay. Schema flexibility was MongoDB’s defining advantage for roughly a decade, and it is precisely the advantage that incumbents and new entrants alike have spent the last several years closing, whether through Postgres’s JSONB support or wire-compatible alternatives like FerretDB. New CEO CJ Desai has framed MongoDB’s next chapter around powering AI-native applications on its existing developer base rather than starting a new go-to-market motion from scratch, a sign that the company recognizes its original wedge needs a successor, not just a defense (Constellation Research, November 2025). Founders building anything with a single, clear technical differentiator should assume that differentiator has a shelf life and plan the next one before the first one erodes, not after.
The third lesson is about succession itself. MongoDB’s leadership transition from Dev Ittycheria to CJ Desai happened from a position of strength, announced alongside preliminary results that beat guidance, not in the middle of a crisis. That sequencing, transitioning leadership when the business is healthy rather than when it is under pressure, gave the new CEO room to set a forward-looking AI strategy rather than spend his first quarters firefighting. Indian founders building toward an eventual leadership handover, whether to a co-founder, a hired executive, or a public-market CEO, should treat the timing of that transition as a strategic decision in its own right, not an afterthought forced by a crisis.
Finally, the international dimension deserves attention. With India having overtaken the United States as the largest source of new open source contributors globally in 2025, the next generation of database decisions, including whether to default to MongoDB, Postgres, or something newer, will increasingly be made by Indian engineering teams first. Companies building developer infrastructure in or for India should treat this audience as a primary, not secondary, distribution channel for exactly the same reasons MongoDB’s open source release worked globally two decades ago.
Expert analysis
The shift in developer sentiment away from MongoDB and back toward relational databases is not a rejection of the original NoSQL thesis, it is closer to convergence. According to Advait Lachake, a software engineer who has written on database tradeoffs, no single database is best in isolation, only the right fit for a specific workload and team, a framing that explains why Postgres and MongoDB increasingly compete less on raw capability and more on which team’s existing skills and workflows fit better (via Nucamp, January 2026). [VISUAL: not applicable to this section] What makes this moment genuinely interesting for analysts is that MongoDB’s revenue keeps growing even as its developer mindshare erodes, suggesting the company’s enterprise and AI-workload customers are stickier and slower to switch than the broader developer population’s stated preferences would imply.
Future outlook
MongoDB’s next phase rests on convincing both new and existing customers that its document model, now extended with full-text search, vector search, and integrated embedding models through its Voyage AI acquisition, is the right foundation for AI-native applications rather than a legacy NoSQL choice from a previous era (StockTitan, March 2026). The company has guided fiscal 2027 revenue to between $2.86 billion and $2.90 billion, implying continued growth in the high teens to low twenties percentage range, even as it competes for AI workload spend against both Postgres-based challengers and other purpose-built vector database vendors. Whether MongoDB can convert its installed enterprise base into the default data layer for AI applications, the same conversion logic that powered the original move to Atlas, will likely determine whether the developer sentiment decline ever shows up in MongoDB’s actual revenue numbers.
The bottom line
MongoDB proved that developers will abandon an entire database paradigm for a better experience, and its own slipping developer mindshare proves that lesson cuts both ways.
Key takeaways
- MongoDB was founded in 2007 as 10gen after its founders hit real scaling limits with relational databases at DoubleClick, and released the MongoDB database as a free, open source project in 2009.
- The company’s commercial breakthrough came not from open source licensing but from Atlas, its 2016 managed cloud product, which now generates 75 percent of total revenue.
- MongoDB closed fiscal 2026 with $2.46 billion in revenue, up 23 percent year over year, and trades on Nasdaq at a market capitalization of roughly $28.5 billion as of June 2026.
- Developer sentiment has reversed since MongoDB’s early NoSQL-era peak, with PostgreSQL now the most-used and most-wanted database while MongoDB has fallen to fifth in want-to-use rankings.
- CJ Desai succeeded Dev Ittycheria as CEO in November 2025, positioning MongoDB’s next chapter around AI-native application infrastructure built on its existing developer base.
- The transferable principle for founders is that an early technical differentiator has a shelf life, and the company that revisits its own monetization and positioning before competitors close the gap is the one that survives the convergence.
- For Indian developer infrastructure builders, the database decisions of a fast-growing domestic developer population represent the same kind of open opportunity MongoDB captured globally two decades ago.
Conclusion
MongoDB’s story is usually told as a clean arc: developers got fed up with SQL, MongoDB gave them something better, and a public company followed. The fuller version is more useful to founders, because it includes the part where the advantage narrows. PostgreSQL’s modern JSON support and serverless variants have closed much of the original gap that made MongoDB feel necessary, and developer sentiment data shows that shift plainly. Yet MongoDB’s revenue keeps growing regardless, because enterprise customers and AI workloads value reliability and existing investment over the latest developer survey results. The real lesson is not that MongoDB won or is now losing, it is that the company has had to earn its relevance twice, once by disrupting SQL and once by adapting fast enough that the disruption it caused does not eventually catch up with it.
TFN Lens
The Schema Pendulum
The Schema Pendulum describes how developer infrastructure preference swings between flexibility, fast iteration, loose schemas, and structure, strong consistency, joins, and governance, as companies and workloads mature, rather than settling permanently on one side.
MongoDB rode the swing toward flexibility for over a decade, then watched PostgreSQL ride the swing back toward structure as JSONB closed the flexibility gap, all while MongoDB’s own revenue kept compounding because enterprise and AI workloads value durability over developer survey sentiment (Vonng, July 2025; MongoDB Investor Relations, March 2026).
The same pendulum logic recurs across the data infrastructure category covered in this cluster, and Indian founders building anything with a database underneath it should plan for the pendulum to swing again rather than assuming today’s developer preference is permanent.
Follow The Founder Nation and NamasteVC for curated startup funding news, grant alerts, and founder stories from India’s startup ecosystem, delivered straight to your feed, every week.
Frequently asked questions
Who founded MongoDB and when? Dwight Merriman, Eliot Horowitz, and Kevin P. Ryan founded the company in 2007 under the name 10gen, after experiencing relational database scaling problems while working together at DoubleClick. The team renamed the company MongoDB, Inc. in 2013 to align it with the database product itself. Merriman and Horowitz had previously served as CTO and lead engineer respectively at DoubleClick, giving the founding team direct, firsthand experience with the scaling problem they set out to solve.
Is MongoDB a public company? Yes, MongoDB has been a publicly traded company on Nasdaq under the ticker MDB since its October 2017 initial public offering. As of mid-June 2026, the company’s market capitalization is approximately $28.5 billion, reflecting fiscal 2026 revenue of $2.46 billion and 23 percent year-over-year growth. The stock has fluctuated significantly over the past year, with a 52-week range between roughly $196 and $445, reflecting how sensitive MongoDB’s valuation is to shifting sentiment around AI infrastructure spending.
How does MongoDB make money? MongoDB primarily generates revenue through MongoDB Atlas, its usage-based managed cloud database service, which represented 75 percent of total revenue by the first quarter of fiscal 2027. The remainder comes from Enterprise Advanced, a self-managed and on-premise licensing product aimed at regulated industries and AI workloads that require data to stay on-premise. The free, open source core database itself generates no direct revenue, it functions as the adoption layer that feeds the paid Atlas funnel.
Is MongoDB profitable? MongoDB remains GAAP unprofitable, reporting a net loss for fiscal 2026 even as total revenue grew 23 percent year over year to $2.46 billion. On a non-GAAP basis, the company reported positive income from operations and net income, reflecting heavy ongoing investment in research and development and go-to-market expansion. The company’s guidance for fiscal 2027 projects continued revenue growth alongside an explicit focus on margin expansion.
What is the difference between MongoDB and PostgreSQL? MongoDB is a document-oriented NoSQL database that stores data in flexible, JSON-like structures without requiring a fixed schema in advance, while PostgreSQL is a relational SQL database built around fixed tables, joins, and strict schema enforcement. PostgreSQL has added increasingly capable JSON support over the past several database releases, narrowing the schema-flexibility gap that originally separated the two approaches. In the 2025 Stack Overflow Developer Survey, PostgreSQL led overall developer usage at 55.6 percent, while MongoDB’s usage sat closer to 24 to 26 percent.
Why did developers originally abandon SQL for MongoDB? Developers moved to MongoDB primarily because relational databases required a rigid, predefined schema that slowed down fast-iterating application development, especially for startups whose data models changed frequently in the early stages of building a product. MongoDB’s document model let engineers store data without committing to a fixed structure upfront, which removed a significant source of friction for teams prioritizing speed over long-term data governance. This advantage was strongest for startups and consumer applications with rapidly evolving requirements, less so for workloads that depended on complex relational joins from the start.
Why are developers now going back to SQL? PostgreSQL’s modern JSONB capabilities now offer much of the schema flexibility that originally made MongoDB attractive, while still preserving ACID transactions, joins, and data integrity constraints that relational databases have always provided. This combination lets teams get MongoDB-style flexibility without giving up the consistency guarantees that become more important as applications mature and data correctness matters more. The 2025 Stack Overflow Developer Survey shows this shift clearly, with MongoDB falling from the most-wanted database between 2017 and 2020 to fifth place by 2025, while PostgreSQL has held the top spot for three consecutive years.
Is MongoDB bigger than PostgreSQL? By revenue and company valuation, yes, since MongoDB is a roughly $28.5 billion public company while PostgreSQL is an open source project with no single owning company or revenue figure of its own. By developer usage and sentiment, no, PostgreSQL leads MongoDB by a wide margin, with 55.6 percent of developers using PostgreSQL in 2025 compared with roughly a quarter of developers using MongoDB. The comparison depends entirely on whether the question is about commercial scale or developer mindshare, and the two metrics currently point in opposite directions.
What can founders learn from MongoDB’s growth strategy? The clearest lesson is that an open source product’s initial commercial model rarely becomes the business that ultimately scales; MongoDB’s pivot from license-and-support revenue to Atlas’s usage-based cloud subscription is what actually built its public-market valuation. Founders should also note that a single technical differentiator, however strong initially, has a shelf life once competitors close the gap, which means the next differentiator needs to be in development before the first one erodes. Finally, MongoDB’s smooth, strength-based leadership transition in 2025 shows the value of planning succession proactively rather than reactively.
What mistakes did MongoDB make along the way? MongoDB’s 2018 shift away from the AGPL license to its own Server Side Public License, which has never received Open Source Initiative approval, created friction with parts of the open source community and directly prompted AWS to build a competing, MongoDB-compatible DocumentDB service. The company also allowed its core technical differentiator, schema flexibility, to remain its primary pitch for years without sufficiently anticipating how quickly PostgreSQL’s JSON capabilities would close that gap. Both decisions left openings that competitors and alternative platforms have since used to chip away at MongoDB’s original developer mindshare.
Sources
- “MongoDB, Inc. Announces Fourth Quarter Fiscal 2026 Financial Results,” MongoDB Investor Relations, March 2, 2026. https://investors.mongodb.com/news-releases/news-release-details/mongodb-inc-announces-fourth-quarter-fiscal-2026-financial
- “MongoDB, Inc. Announces First Quarter Fiscal 2027 Financial Results,” MongoDB Investor Relations, May 28, 2026. https://www.sec.gov/Archives/edgar/data/0001441816/000162828026038798/mdb-043026xex991xrelease.htm
- “MongoDB Announces Leadership Transition,” MongoDB Investor Relations, November 3, 2025. https://investors.mongodb.com/news-releases/news-release-details/mongodb-announces-leadership-transition
- “Our Story,” MongoDB. https://www.mongodb.com/company/our-story
- “MDB: MongoDB Inc, Stock Price, Quote and News,” CNBC, June 2026. https://www.cnbc.com/quotes/MDB
- “10gen Is Now MongoDB To Reflect Focus On NoSQL Database,” TechCrunch, August 27, 2013. https://techcrunch.com/2013/08/27/10gen-is-now-mongodb-to-reflect-focus-on-nosql-database/amp/
- “PostgreSQL Has Dominated the Database World,” Vonng, July 31, 2025. https://vonng.com/en/pg/so2025-pg/
- “MongoDB vs PostgreSQL in 2026: NoSQL vs SQL for Full Stack Apps,” Nucamp, January 19, 2026. https://www.nucamp.co/blog/mongodb-vs-postgresql-in-2026-nosql-vs-sql-for-full-stack-apps
- “MongoDB names CJ Desai CEO,” Constellation Research, November 3, 2025. https://www.constellationr.com/insights/news/mongodb-names-cj-desai-ceo
- “MongoDB co-creator explains why ‘NoSQL’ came to be, and why open source mastery is an elusive goal,” Medium, January 7, 2016. https://medium.com/s-c-a-l-e/mongodb-co-creator-explains-why-nosql-came-to-be-and-why-open-source-mastery-is-an-elusive-goal-3a138480b9cd
© 2026 The Founder Nation. All rights reserved.
