MMS • RSS
MongoDB’s stock (NASDAQ:MDB) has almost doubled in value this year, recovering from its lows in late 2022 alongside momentum in the broader tech sector. Investor optimism that the most aggressive rate hike cycle in decades is nearing an end, and surging AI interest have been this year’s market-driving themes. And MongoDB and its software peers have been key beneficiaries of the combination this year.
However, we remain cautious of the premium that has been allocated to the stock at current levels. MongoDB management appears to be one of the more conservative and transparent among large-cap high-growth software names within our coverage – and we think credit is warranted to this prudence. Despite the strong beat and raise in fiscal Q2, much of the latest earnings call’s airtime have been dedicated to reminding investors that the upward-adjusted guidance is the result of a one-time outperformance in non-Atlas sales that is not expected to repeat in the second half.
In the following analysis, we will provide an overview of how MongoDB differentiates itself from legacy databases dominated by juggernauts like Oracle (ORCL), which remain the predominant choice in the industry. Then, we will discuss the ensuing tailwinds to MongoDB’s offering and strategy from accelerating AI and broader cloud adoption, as well as the near-term fundamental considerations.
We believe the innovative work that MongoDB is doing to improve efficiency, reliance and scalability of databases for developers and enterprise users will benefit from the increasing cloud- and AI-first era. However, the stock’s current valuation premium risks exposure to weakness in the near-term given the unfavourable combination of persistent macroeconomic, company-specific and demand environment uncertainties.
Relational vs. NoSQL Databases
The advent of connectivity has created massive troves of data that drive decision-making processes every day. This has made databases an increasingly critical component of the enterprise IT structure. Databases are typically categorized in two types today – 1) relational, and 2) NoSQL databases.
Relational databases store data in a tabled format. It typically requires a fixed structure, or “schema”, of data in rows and columns. Relational databases often use SQL to manage and support “creating, reading, updating, and deleting (“CRUD”) operations, and “atomicity, consistency, isolation, and durability” (“ACID”) transactions on data elements.
Relational databases are viewed as the legacy offering, which has long been the industry standard. However, increasingly complex workloads due to massive data troves are unveiling the cost and scalability inefficiencies of relational databases.
The table structure of stored data in relational databases often contain basic information between a heading and value. Take user data for instance (example taken from mongodb.com) – each heading (ID, first name, email, cell) is allocated a value (1, Tom, email@example.com, 765-555-5555) in a table structure.
An additional heading “businesses” can be added to describe Tom’s businesses. But what if there is more than one business or additional data pertaining to each business? In this case, a new table can be created to document info on Tom’s businesses:
In order to bridge the two tables of data (i.e. “businesses” and “users”), a “foreign key” is used to “reference the ID column in the users table”. A similar table extension structure is used to document all data relevant to Tom (e.g. hobbies, businesses, etc.) in a relational database.
Given increasingly complex and high-volume workloads in today’s data environment, organizing highly structured data and building foreign keys to operate a relational database can become an expensive and inefficient endeavour to operate. While relational databases remain a critical function to enterprise IT environments, it is gradually becoming obsolete in supporting the cloud- and AI-first era due to the technology’s inherently low productivity yield relative to NoSQL databases, such as MongoDB’s Atlas.
Contrary to relational databases, NoSQL (or not only SQL / non-SQL) databases are not limited to table structures. NoSQL databases can store “unstructured or semi-structured data”, which makes them “schema-agnostic”. This effectively addresses inefficiencies such as having to allocate substantial resources toward structuring data and linking tables. It also improves the costs of deployment, as well as developer productivity, making NoSQL databases a more scalable option. There are four common types of NoSQL databases – 1) document store, 2) key-value store, 3) wide-column store, and 4) graph store. MongoDB’s NoSQL database adopts a document-based architecture.
Document Store – Circling back to the earlier data sample on Tom, a document store effectively eliminates the need for structuring data into tabled rows and columns with linking keys. Instead, all of Tom’s data is stored in a single document structured by headers and values in a hierarchical structure (kind of like a word doc in bullet-point format). A document database can store a collection of documents with similar content (e.g. information on different users in addition to Tom). Each document can also be structured differently, given there is no requirement for a fixed schema. However, some document stores offer “schema validation”, which provides data users with the option to lock-in a specific schema for structuring the data.
Which is MongoDB Database?
MongoDB operates a NoSQL document-based architecture database. The company embraces a “run anywhere” strategy, where MongoDB can be “fully managed as-a-service or self-managed in the cloud, on-premise, or in a hybrid environment”. This essentially allows customers to run MongoDB in whichever environment that best serves their needs.
MongoDB Atlas – Atlas is MongoDB’s fully-managed database-as-a-service (“DBaaS”) offering. Atlas is a one-stop-shop NoSQL database, facilitating user data infrastructures from provisioning to set-up to deployment. Atlas is the only independent software vendor available on all of AWS (AMZN), Azure (MSFT), and Google Cloud Platform (GOOG / GOOGL), making it a convenient and accessible choice for customers no matter who their primary cloud service provider is. Customers can conveniently choose the cloud service provider, as well as respective configurations (e.g. region, instance size, etc.) they want to run Atlas on and start building on the NoSQL database. Atlas offers a comprehensive range of data features, such as Atlas Search Overview, device sync and MongoDB Atlas Charts to ensure data accessibility, availability and visualization for users. The latest features include Atlas Vector Search, Atlas Stream Processing, and Relational Migrator to better address modern data and application needs.
MongoDB Enterprise Advanced – EA is a self-managed database for enterprise customers that can be “run in the cloud, on-premise or in a hybrid environment”. EA subscriptions give customers access to features such as the following:
- MongoDB Enterprise Server– a commercial-focused database designed to address enterprise-specific data needs, such as security, audit functionality, authentication and authorization.
- Enterprise Management Capabilities – EA customers can choose between “Cloud Manager Premium” and “Ops Manager” for managing their data within MongoDB. Cloud Manager Premium is for self-managed customers looking to run MongoDB in the cloud. Ops Manager is for self-managed customers looking to deploy the database on-premise.
- Analytics Integrations – Data users can integrate their “existing business intelligence and analytics tools” to data within the MongoDB platform to efficiently and seamlessly drive relevant insight.
In addition to subscriptions, MongoDB also generates revenue from the sale of professional services, such as consulting and training services for customers deploying Atlas and/or EA. The company also offers free services as part of its go-to-market strategy. These include the free-to-download Community Server database and a “freemium” Atlas tier to drive brand and product awareness.
Implications of the Cloud- and AI-First Era for NoSQL Databases
Productivity gains and scalability are primary advantages of NoSQL databases. As mentioned in the earlier section, workloads are becoming increasingly massive and complex in the cloud- and AI-first environment, which drives up developer costs even when data storage costs have already become cheaper with scale. The flexible schema of NoSQL databases allow processing of large workloads at scale and reduced latency, while the data retrieval process can also be simplified via integrated features such as Atlas Search and Stream Processing. Since the emergence of NoSQL databases in the late 2000s, costs of managing data have come down significantly due to the resulting productivity gains.
In MongoDB’s case, its document-based architecture NoSQL database allows users to store unstructured / semi-structured data in a flexible schema. This enables the company to support a “broader range of use cases” and, inadvertently, penetrate a larger TAM. NoSQL opportunities are expected to represent $35 billion of the anticipated $150 billion market for global database software, highlighting the secular tailwinds ahead for MongoDB. Combined with the comprehensive set of features such as “indexing support and replication” integrated into Atlas and EA, MongoDB enables massive scalability to support next-generation data needs and applications, such as the deployment of AI technologies.
While it is not expected that relational databases will face a secular decline anytime soon, since they remain “the best option for certain use cases”, NoSQL system vendors such as MongoDB are expected to be emerging share gainers by facilitating the data requirements of “modern applications” such as AI deployments at scale. And the company has deployed a unique go-to-market strategy that addresses the needs of not only data scientists but also developers, as observed through “developer-centric” Atlas innovations such as Vendor Search and Stream Processing. This has effectively reinforced MongoDB’s appeal to end-users and enabled the company to consistently take market share from legacy relational database systems.
Stepping closer to near-term considerations over MongoDB’s operating environment, we remain cautious of the looming macroeconomic impacts on Atlas adoption as management had repeatedly warned about. Atlas is the biggest source of revenue at MongoDB, representing more than 60% of its consolidated sales with a “$1 billion-plus revenue run-rate” exiting fiscal Q2. Weakness on this front could potentially impact MongoDB’s near-term growth outlook, which management has prudently sought to temper at the latest earnings call. Yet, the stock’s lofty valuation premium at current levels has remained resilient, though we are hesitant about its durability through the remainder of the year.
Atlas runs a consumption-based pricing model, where customers are only billed for the workloads that go through the platform. Recall from our recent discussion on consumption-based pricing models – they are typically more prone to macroeconomic impacts relative to subscription models, since they offer customers the flexibility to scale up/scale down usage. During an uncertain spending environment, consumption-based revenue are typically first to go, while subscription revenues stay relatively resilient. This is corroborated by a sequential moderation averaging 900 bps in Atlas revenue growth over the past 12 months. And management expects the continuation of similar macroeconomic impacts on Atlas consumption heading into the second half of fiscal 2024.
However, we expect continued product innovation (e.g. Relational Migrator; Vector Search; Stream Processing) and a higher starting ARR for fiscal 2H24 coming off of fiscal Q2 outperformance to be compensatory factors to consumption and revenue growth amid near-term cyclical challenges. Meanwhile, MongoDB’s strategic shift towards greater reliance on indirect sales via hyperscaler marketplaces also paves the way to achieving operating leverage once cyclical tailwinds return. As mentioned in the earlier section, MongoDB is currently the only ISV available on all three of the largest cloud service providers. The company has experienced increasing volumes of customer self-serve Atlas deployments via AWS, GCP and Azure marketplaces. Self-serve customers pay for Atlas by “drawing down their existing cloud commitments”, which minimizes friction to the consumption process. This has effectively reduced direct sales costs, while also maintaining acquisition of incremental workloads by leveraging exposure and convenience of its hyperscaler go-to-market partners.
We forecast Atlas revenues finish fiscal 2024 with 32% y/y growth, and account for 66% of consolidated sales.
Our longer-term projections for Atlas estimates expansion at a five-year CAGR of about 22%. The assumption is consistent with historical consumption trends, and considers the product’s rate of ARR expansion and significant TAM supported by secular demands discussed in the foregoing analysis.
On the non-Atlas front, EA consumption will remain the biggest driver, though we are cautious of fiscal Q2 outperformance being a one-time positive impact on the product’s expanded sales mix contribution. We are forecasting a double-digit sequential decline in non-Atlas revenue for fiscal Q3, with moderation into fiscal Q4 to normalize for stronger-than-expected Q2 outperformance.
We expect the product category to expand sales at a five-year CAGR of 12% through fiscal 2028, and gradually reduce its sales mix contribution to about a quarter over the forecast period. The assumptions applied are consistent with historical EA consumption trends, as well as expectations for the increasing secular shift from on-premise to the cloud.
Taken together with nominal services revenue (~3% of sales mix), we expect total MongoDB revenue to expand by 26% y/y in fiscal 2024. Our forecasts estimates reacceleration in fiscal 2026 when cyclical headwinds to the consumption model dissipates and NoSQL database adoption gains momentum on the back of increasingly modern applications.
Meanwhile, profit margins are expected to normalize from fiscal Q2 outperformance back to the lower-end through fiscal 2H24, in line with the slower pace of growth. Continued operating leverage, bolstered by improved efficiencies in the go-to-market strategy is expected to drive GAAP profitability exiting fiscal 2026.
By applying the discounted cash flow (“DCF”) valuation approach to our fundamental projections over the five-year forecast period, alongside a 9.7% WACC based on MongoDB’s capital structure and risk profile, the stock’s performance at current levels ($371, September 13 close) implies a 6.7% perpetual growth rate, or 25x terminal value multiple.
This represents a significant valuation premium, especially considering the shaky fundamental backdrop amid the uncertain enterprise spending environment. Paired with uncertainties to the elevated rate environment, which risks further discounting the value of MongoDB’s future cash flows, we expect some volatility to the stock at current levels.
However, from a longer-term perspective ex-macros, the stock’s current price is adequately reflective of its longer-term growth trajectory that is supported by robust secular tailwinds and reinforced by emerging AI momentum. This is in line with our extended DCF analysis based on fundamental projections over a 10-year forecast period, in line with consensus growth expectations.
By maintaining the 9.7% WACC to reflect MongoDB’s capital structure, and applying a 1.5% estimated perpetual growth rate in line with the pace of economic expansion typically used to value maturing companies with stable cash flows, the company yields an estimated intrinsic value of $31 billion. This represents upside potential of 17% from its current market value of $26.5 billion. But with the anticipated realization of share dilution at 83 million units over the extended forecast period, the stock price remains at $372, in line with its currently traded levels.
This is further corroborated by the multiple-based valuation approach, which considers our near-term base case fundamental forecast for MongoDB relative to its software peers.
Considering our base case growth projections for the company, as well as MongoDB’s current valuation curve (dotted red line) relative to peers (dotted blue line), the stock should trade at approximately $353 apiece based on the current 71 million shares outstanding in the market.
We believe the average outcome of the two valuation methods of $363 is an adequate representation of our base case price target for MongoDB over the next 12 months, as it captures both near- and longer-term fundamental considerations.
The Bottom Line
We believe MongoDB’s increasing reach in the NoSQL database market, bolstered by its continued product innovations and unique go-to-market strategy in the field, is well-positioned for capturing longer-term AI and cloud tailwinds already underway. A valuation premium is warranted for the stock, given the favourable secular backdrop and significant TAM that has yet to fully materialize. However, considering risks of near-term weakness to its fundamentals – especially given Atlas’ consumption-based model – and an elevated rate environment, we remain cautious on the durability of MongoDB’s valuation premium at current market levels. We expect some volatility to the stock in the next 12 months, driven by the elevated rate environment for valuations and uncertain demand environment for underlying business fundamentals. This could potentially open up a compelling entry opportunity at the lower $300-range.