MDB’s AI Potential: Growth Ahead, but Valuation Leaves No Room for Error

MDB 0.00%↑ has emerged as a dominant player in the database market, carving out a niche for itself by offering a flexible, scalable alternative to traditional relational databases. The company’s core offering is a document-based NoSQL database, which is a critical component for developers looking to build modern, data-driven applications. As companies move away from rigid data infrastructures and lean into cloud-native and AI-driven architectures, MongoDB is positioning itself to capture a growing share of an $80 billion industry.

In this post, we'll dive into MongoDB’s unique value proposition, the pros and cons of its approach compared to traditional relational databases, and the growth levers that are driving the company’s future.

What Sets MDB Apart

To understand MDB's success, it's important to grasp what differentiates it from traditional databases. Relational databases, long the standard for data management, organize information into structured tables with rows and columns. This works well for highly structured data but struggles with the flexibility and scale demanded by modern applications like real-time analytics, AI models, and internet-scale applications. In contrast, MDB uses a document-based model, where data is stored in flexible, JSON-like documents. This allows developers to handle diverse data types more easily and provides agility, which is critical for rapidly evolving applications. As Michael Gordon, MDB's COO and CFO, noted at a recent conference, the flexibility and scalability of MDB make it a perfect fit for companies modernizing their data infrastructure.

Further expanding this advantage is Atlas, MDB's fully-managed cloud database platform. Atlas abstracts the complexity of database management, allowing companies to deploy MDB's databases across AWS, Azure, and Google Cloud without friction. Atlas has now grown to represent over 70% of MDB's revenue, driven by the shift to the cloud. With Atlas, businesses can scale without worrying about infrastructure management, a key driver of its rapid adoption and continued outperformance relative to other areas of the company.

Why Developers Love MDB

A major reason for MDB's strong growth is its close alignment with the needs of developers. MDB was built with the developer experience in mind, and its document model is intuitive and easy to use, a critical factor in gaining traction with development teams. Developers, after all, are the ones who ultimately decide which databases power their applications. Traditional relational databases like Oracle and SQL require developers to work within a rigid structure, which can slow development cycles and add complexity. MDB, on the other hand, is designed to handle unstructured and semi-structured data, enabling developers to iterate quickly and release features faster. This is why MDB has become especially popular with startups and companies building fast-moving, cloud-native applications.

As Gordon mentioned, MDB's ability to "win the hearts and minds of developers" has been a core driver of its success.

MDB offers several technical advantages over traditional relational databases:

• Scalability and Flexibility: MDB's document-based model allows for horizontal scaling, meaning it can handle massive amounts of data by adding more servers. In contrast, relational databases often require vertical scaling, which can become costly and complex as applications grow.

• Dynamic Schemas: With MDB, developers don’t need to pre-define schemas, giving them the freedom to work with evolving datasets. Relational databases, by contrast, have rigid schemas that need to be defined upfront, making changes or iterations expensive and time-consuming.

• Handling Unstructured Data: MDB excels at managing unstructured or semi-structured data. As applications collect data from diverse sources—like IoT devices, social media, or AI systems—MDB's flexible document model is better suited to store this data without extensive pre-processing.

However, this flexibility comes with trade-offs. For industries where transactional integrity is critical, such as finance, relational databases still reign supreme. Relational databases excel at handling complex, multi-row ACID (Atomicity, Consistency, Isolation, Durability) transactions, which ensures that each transaction is processed reliably, even in the event of a system failure. While MDB has added support for multi-document ACID transactions, its performance in this area still lags behind more mature relational systems.

Despite its strengths, MDB faces several challenges, especially when competing against legacy relational systems optimized over decades. Here are some areas where MDB may struggle:

• Migration Costs: Moving from a relational database to MDB is not a simple task. The two database types have fundamentally different architectures, meaning that organizations may need to rearchitect their applications to fully leverage MDB's capabilities. This can make migrations expensive and time-consuming.

• Transactional Complexity: Relational databases like Oracle or MySQL have decades of optimizations that make them better suited for handling complex, high-volume transactional workloads. MDB's handling of multi-document ACID transactions is still maturing, which could limit its use in specific industries.

• Enterprise Reluctance: Large enterprises entrenched in legacy systems may be reluctant to migrate away from relational databases unless there's a clear and significant value proposition. As MDB has acknowledged, moving applications from relational to NoSQL is difficult work, requiring clear business value to undertake such efforts.

AI and Cloud as a Growth Lever for MDB

Cloud adoption continues to be a major growth driver for MDB, and Atlas is at the center of this shift. As more enterprises migrate their workloads to the cloud, they are looking for database solutions that can scale with their needs. Atlas simplifies the management of databases across multiple cloud providers, making it an attractive choice for companies with hybrid or multi-cloud strategies. MDB’s leadership expects Atlas to be a cornerstone of their growth in the coming years, as it captures a larger share of the $80 billion database market.

Another significant growth lever is the rise of AI-driven applications. MDB is positioning itself as a key player in the AI space by enabling AI and machine learning models to process large, flexible datasets. MDB's document model is particularly suited to handle the unstructured data that AI systems often require. They are investing heavily in features like vector search, which allows AI models to quickly search through and retrieve relevant data points. This is particularly useful for applications like chatbots, recommendation engines, and real-time analytics. As Gordon highlighted, the company sees generative AI as a tailwind that will drive more applications—and thus more database usage—in the coming years.

AI-driven applications are becoming increasingly common across industries, whether it’s for natural language processing , recommendation systems, real-time analytics, or intelligent automation. These applications require databases that can handle large, diverse, and unstructured datasets, often at scale. This is where MDB’s flexible, document-based architecture shines. MDB’s document model allows for seamless handling of unstructured data, which is typically used in AI and machine learning models. Unlike relational databases, which struggle with the complexity and diversity of AI datasets, MDB’s schema-less structure makes it easier for developers to store, retrieve, and manipulate large volumes of AI-relevant data. At the September 2024 conference, MDB executives highlighted that the flexibility of its database architecture positions it well to handle the influx of AI applications. AI's ability to generate more software and data leads to more demand for scalable, flexible databases—an area where MDB thrives.

MDB has long seen the migration of legacy relational databases as a key opportunity. However, this is a slow-moving process, as replatforming mission-critical applications is not something businesses take lightly. One of the major hurdles in transitioning from traditional relational databases to modern databases is the complexity and cost of migration. AI can help simplify this process, potentially unlocking a significant new revenue stream for MDB. Generative AI, particularly tools designed to automate migration, can reduce the technical debt associated with moving from legacy SQL-based systems to MDB. This could accelerate the pace at which enterprises re-platform their databases, as AI tools could automate much of the migration work that currently requires extensive human intervention. As Gordon noted during the company's presentation, AI offers an opportunity to "shrink the amount of time to reduce the cost and complexity of migrating" traditional relational databases to MDB’s platform​. By lowering the barriers to entry, AI could unlock a massive opportunity for MongoDB to gain market share from legacy databases.

AI tools like code generation and auto-completion systems (e.g., GitHub Copilot, AWS CodeWhisperer) will make developers more productive, enabling them to build more applications faster. This increase in the number of applications built translates directly into increased demand for databases, as every application requires a backend system to store and manage data. MDB is expected to benefit from this productivity surge. Its popularity among developers, thanks to its flexibility and ease of use, positions it to be the go-to database for many of the new applications that will be built using AI. The company has been very clear that as developers use AI to generate code and build applications faster, the overall demand for databases will grow. This market expansion will likely play in MDB's favor, as many of these new applications will require the flexibility that its platform offers.

MongoDB has been investing in AI-related features, such as vector search capabilities. Vector search is crucial for AI-driven applications, such as recommendation engines and large language models because it allows databases to index and query complex, high-dimensional data, like embeddings produced by machine learning models. For example, vector search is a key component in building AI models that need to quickly identify patterns or retrieve relevant data from vast datasets. By integrating these AI-driven features into its core offering, MDB can capture more AI and ML workloads that require advanced data management and retrieval capabilities. Serge Tanjga, MongoDB’s Vice President of Finance, remarked on the strategic importance of such AI-driven features during the company’s conference, noting that features like vector search and other AI-ready capabilities are critical to winning modern workloads.

AI models, especially those trained on large datasets, generate significant database activity, from reading and writing data to querying vast amounts of information in real-time. Atlas, MDB’s fully managed cloud database service, is particularly well-positioned to capture these workloads. As companies move their AI workloads to the cloud, they will increasingly look for scalable, flexible databases that can handle large volumes of unstructured data and operate seamlessly in multi-cloud environments. Atlas allows enterprises to do this without having to worry about managing the infrastructure themselves. This creates a direct path for monetizing the surge in AI-driven data workloads. Traditional relational databases, such as those offered by Oracle and Microsoft, were not designed for the high-dimensional, unstructured data typical of AI workloads. As a result, these legacy systems often require extensive customization or workarounds to handle AI applications. Atlas, by contrast, is well-suited to handle the unique demands of AI, giving it a competitive advantage in capturing the next wave of database adoption. The company’s focus on AI-friendly features, like its document model and search capabilities, positions it to be a leading player in the AI space, while relational databases may struggle to adapt to the new paradigm. As noted by executives, AI presents a "market share opportunity" as more enterprises realize they need modern databases to fully leverage the benefits of AI​.

Risks AI Poses to MDB

AI presents significant growth opportunities, but it also introduces several risks that could impact the company’s operations, competitive positioning, and long-term profitability. As AI-driven applications grow in prominence, specialized AI databases and tools tailored for these workloads are becoming more common. New database technologies designed specifically for ML and AI workloads could compete directly with MDB's offerings. Vector databases like Pinecone or Weaviate, for instance, are designed to handle the specific needs of AI applications by efficiently storing and searching through high-dimensional vectors. These technologies, built from the ground up for AI, could capture a portion of the AI-driven database market, making it harder for general-purpose databases to maintain or expand their share in this rapidly growing niche. MDB has started to integrate AI-driven features like vector search, but the company's general-purpose database architecture might not be as optimized for AI-specific workloads as some of these new, specialized competitors.

Cloud hyperscalers like AWS, Google Cloud, and Azure are heavily investing in AI and machine learning capabilities, and they are incorporating AI-ready database solutions within their platforms. These hyperscalers often bundle their proprietary databases with their AI and cloud offerings, which could pose a serious threat to Atlas. As AI workloads increase, there is a growing risk that enterprises will gravitate toward using integrated AI platforms from these hyperscalers, which offer tighter integration between data storage, model training, and AI inference. For example, AWS offers its own AI-optimized databases like Aurora (relational) and DynamoDB (NoSQL), and Google Cloud has BigQuery, which is increasingly optimized for AI workloads. MDB’s platform-agnostic nature may not be enough to deter customers from opting for these bundled, often more cost-effective solutions​.

In the AI ecosystem, open-source frameworks and tools are widely adopted due to their flexibility and cost-effectiveness. Open-source AI databases and platforms, such as PostgreSQL (with AI extensions) or the aforementioned vector-specific solutions, could siphon market share, especially among cost-conscious developers and startups. Open-source tools are often appealing because they come without the licensing fees associated with proprietary software, and they offer developers greater control over customization. This trend could pose a risk to MDB’s premium pricing model and the potential for customer churn if organizations decide to use free, open-source alternatives instead of paying for MDB’s services. Moreover, some open-source AI solutions have strong community support, which helps them evolve rapidly, reducing MDB's competitive advantage.

AI-driven applications are highly resource-intensive, both in terms of compute and data storage. As companies scale their AI workloads, they will demand more from their database providers in terms of performance, scalability, and low-latency data retrieval. While MDB has invested in scaling its infrastructure to meet these demands, the rapid pace of AI adoption could outstrip its ability to keep up. For example, AI models often need to store vast amounts of high-dimensional data, which can strain traditional database infrastructures. If they cannot efficiently manage these resource-intensive workloads, it risks falling behind other database providers or requiring significant capital expenditures to upgrade its infrastructure, which could negatively impact margins.

AI could fundamentally change how applications are developed. As AI-assisted development tools, like code generators and auto-completion software, become more widespread, the role of developers in choosing databases may diminish. Currently, MDB enjoys strong developer affinity due to its ease of use and flexibility, but if AI tools start recommending or even automatically selecting databases optimized for specific use cases, MDB might lose its edge. AI tools could favor databases that are highly optimized for AI-specific workloads, driving developers toward more specialized solutions. This would undercut one of their key advantages—its popularity among developers—and shift purchasing decisions towards tools that are tightly integrated with AI development ecosystems.

As AI progresses, it is becoming capable of automating many traditional database management tasks, such as optimization, indexing, and schema design. This trend could commoditize the value of the underlying database, making the choice of database less important for some AI-driven applications. As databases become more self-optimizing, the key differentiator for customers may shift from database performance to other factors like integration with AI tools and overall cost. This presents a long-term risk if AI diminishes the perceived value of its database services. Automation might erode some of the technological advantages MDB offers today, making it harder for the company to justify its premium pricing, especially if competitors can deliver similar automation features at a lower cost.

While AI represents a significant growth opportunity, it also introduces several risks that could disrupt the company’s trajectory. The rise of specialized AI databases, competition from hyperscalers, and the increasing popularity of open-source alternatives all pose competitive threats. Additionally, AI could shift customer preferences away from MDB, particularly if automation reduces the importance of developer-driven decision-making or leads to unanticipated pricing pressures. To navigate these risks, MDB will need to continue innovating—both in terms of AI capabilities and infrastructure scalability—while ensuring that its platform remains the database of choice for developers building the next generation of AI applications. Success will require careful management of AI-related risks, strong partnerships, and continued investment in both technology and customer trust.

MDB's AI Features and Competitive Edge

MongoDB’s fully managed DBaaS, offers several features that specifically cater to AI and ML workloads. As AI applications grow in complexity and scope, Atlas has integrated advanced tools and capabilities to ensure developers can efficiently store, query, and manage the vast amounts of data AI-driven systems require. Atlas has evolved into a platform well-suited for AI and machine learning workloads, with features like vector search, real-time analytics, stream processing, and serverless scaling that cater to the demands of modern AI applications. Its ability to integrate with popular AI ecosystems, handle large datasets, and provide flexible and scalable infrastructure makes it an attractive choice for developers building AI-driven systems.

1. Vector Search

One of the most important features for AI workloads is vector search. Vector search is essential for applications like recommendation engines, image and document retrieval, and LLMs. These applications rely on embeddings or vectors—high-dimensional numerical representations of data such as words, images, or customer behavior patterns. In traditional databases, searching through these vectors would be computationally intensive and inefficient. However, Atlas has integrated vector search capabilities, enabling AI models to quickly find and retrieve relevant data points from vast datasets. This is particularly useful in AI use cases that involve semantic search, similarity matching, and other complex data queries where conventional keyword-based searches fall short. At the company's recent conference, executives emphasized that vector search is becoming a core requirement for AI applications, and its implementation in Atlas is designed to help developers seamlessly integrate advanced search capabilities into their applications​

2. Real-Time Analytics and Stream Processing

AI applications, especially those dealing with real-time data, require the ability to process and analyze information as it’s ingested. Atlas has integrated real-time analytics and stream processing capabilities that allow developers to analyze data in motion, making it ideal for AI use cases like fraud detection, monitoring systems, and dynamic recommendation engines. Atlas' stream processing enables users to handle event-driven architectures, where the system can trigger AI models based on real-time data inputs, such as sensor readings, transactional data, or social media feeds. This feature is crucial for AI applications that need to react immediately to changing data patterns or user behaviors.

3. Atlas Search

Beyond vector search, MongoDB also offers Atlas Search, which is built on the Apache Lucene engine. This feature allows for more sophisticated search functionalities, such as full-text search, faceted search, and filtering, making it easier to build AI-powered applications that involve complex search queries and user interactions. AI applications often need to process and search through large amounts of textual data, and Atlas Search allows developers to implement these capabilities without relying on external search platforms. The tight integration between MDB and Atlas Search ensures that users can query, index, and analyze data without having to leave the ecosystem.

4. Serverless Architecture for AI-Driven Workloads

With Atlas' serverless architecture, AI developers can build applications without worrying about the underlying infrastructure. This is particularly beneficial for AI workloads that may have unpredictable or spiky demands. Serverless architecture automatically scales up or down based on demand, making it cost-effective for AI-driven applications that might require high processing power at certain times but low usage at others. For AI developers, this means that Atlas can handle large-scale training or inference workloads when needed, while also optimizing for cost by scaling down during idle times. This flexibility makes it easier to manage AI workloads without overprovisioning infrastructure, which can be a common issue with traditional hosting.

6. Data Lake for AI Workloads

Atlas provides a Data Lake service that allows for the storage and querying of massive, unstructured datasets—often a key requirement for AI applications. AI models, especially those that involve deep learning, need access to large datasets to train effectively, and MDB’s Data Lake can store these large datasets while still allowing for efficient querying. Atlas Data Lake can natively query data in various formats (e.g., JSON, CSV, Avro) and integrates with other features like Atlas Search and aggregation pipelines, which is useful for processing large volumes of data typically used in AI models. This allows developers to build complex AI-driven systems without worrying about data scalability or storage limitations.

7. Flexible Schema for AI Training Data

One of the strongest selling points is its schema flexibility, which is especially valuable for AI and ML applications. AI workloads typically involve large, evolving datasets, and MDB’s schema-less architecture allows developers to quickly modify and expand their data models as new data is ingested. This is particularly important in AI development, where models often require continuous retraining with new data. Their flexible data model can store diverse types of data, such as images, text, or numerical data, which are often needed for training AI models. Additionally, its dynamic schema makes it easier to accommodate changes in data structure, which is crucial for evolving AI systems that rely on continuously updated data inputs.

8. Horizontal Scaling Through Sharding

Atlas excels in horizontal scaling via sharding, which is the process of distributing data across multiple servers. When a database reaches capacity limits, Atlas automatically scales by distributing the load across additional shards (clusters of servers). This makes it possible to handle immense datasets and high query volumes without hitting performance bottlenecks. Sharding is particularly useful for applications with growing datasets, such as AI workloads, real-time analytics, or e-commerce platforms where traffic can increase rapidly. By distributing data across multiple machines, Atlas ensures that the system remains performant and responsive as the amount of data grows.

9. Multi-Cloud and Multi-Region Scaling

Atlas is multi-cloud by design, supporting seamless deployment across AWS, Google Cloud, and Microsoft Azure. This multi-cloud capability allows companies to scale across different cloud environments based on their specific needs, ensuring redundancy, flexibility, and performance optimization. In addition, Atlas allows for multi-region deployments, meaning users can replicate data across various geographic regions. This ensures low-latency access to data for global applications, disaster recovery in case of regional outages, and compliance with data residency requirements. Multi-cloud and multi-region scaling are crucial for enterprises that need to serve users globally or ensure high availability in mission-critical applications, such as financial services, e-commerce, or social media platforms.

10. Scalability in Pricing

Finally, the pay-as-you-go pricing model in Atlas aligns costs directly with the level of scaling. As businesses scale up their workloads, they can manage their costs effectively without incurring unnecessary expenses. This pricing structure makes Atlas a scalable solution from both a technical and financial perspective, which is particularly attractive to startups and growing businesses looking to manage their resources effectively as they expand.

Bullish/Bearish Case for MDB

• Atlas Growth: Atlas platform continues to be the company’s primary growth engine. With enterprises rapidly adopting cloud infrastructure, Atlas provides a way for MongoDB to generate recurring, predictable revenue while expanding its customer base.

• AI Tailwinds: As AI becomes a more prominent driver of enterprise software development, MDB’s ability to handle unstructured data gives it an advantage over more rigid relational databases. The rise of AI-driven applications will likely create more demand for MDB’s flexible, scalable database architecture.

• Developer-First Approach: MDB’s success is largely driven by its deep integration with the developer community. Developers love MDB because it simplifies their work and allows them to build applications faster. This developer-first approach has created strong mindshare, and as more companies invest in custom-built software, is well-positioned to capitalize on this trend.

• Relational Database Competition: Despite MDB’s strengths, relational databases still dominate the market. Many enterprises remain reluctant to switch from these legacy systems, and MDB’s migration tools may not be compelling enough to disrupt these entrenched players at scale.

• Macro Sensitivity: Like many tech companies, MDB’s growth is tied to broader economic conditions. Recent earnings calls have highlighted that slower macroeconomic conditions have dampened the growth of existing workloads on Atlas, and any prolonged downturn could weigh on future performance.

• Valuation Risks: MDB’s stock has consistently traded at high multiples relative to its revenue and FCF (or lack thereof). For some investors, this raises concerns that the company’s growth may not be able to justify its lofty valuation, particularly if it experiences any significant slowdown, especially driven my increased competition.

Valuation

Sell-side consensus estimates project revenue to grow at a 22% CAGR, expanding from $1.6 billion in CY2024 to just under $4.1 billion by CY2028. FCF margins are expected to improve from 6% in 2024 to 16% by 2028. At today’s EV of around $21 billion, MongoDB is trading at 10.5x EV/2024 sales, with topline growth projected at 15% in 2024, accelerating to 17% in 2025 and 22% in 2026. This is far from a bargain, and it’s difficult to argue that current valuation offers any immediate upside without flawless execution.

However, a case can be made for a more bullish scenario. If MDB can capture an increasing share of AI workloads while scaling FCF margins through operating leverage, the upside becomes more compelling. I’ve modeled a scenario where revenue grows at a 26% CAGR to $5 billion by 2028—$1 billion higher than consensus estimates—with FCF margins reaching 25%, 10 percentage points above the street's estimates. In this scenario, MDB would produce $1.25 billion in FCF by 2028. For investors to see a 10% annualized return over the next four years, the stock would need to trade at 30x FCF—a $37 billion EV, or about $440 per share, assuming 2% annual dilution. Given MDB’s expected low-20s topline growth exit rate by 2028, 30x FCF wouldn’t be an exorbitant multiple, but this scenario does require MongoDB to hit it out of the park both in business execution and in leveraging its operating model.

On the flip side, if we assume the current consensus scenario plays out—where revenue reaches $4.1 billion and FCF margins improve to 18%—MongoDB would need to trade at nearly 50x FCF in 2028 to generate a double-digit annualized return. Achieving this multiple is hard to justify, especially when considering that a large portion of their FCF is driven by SBC, which lowers the quality of its cash flow. At today’s valuation, there is no room for error, and MDB faces several obstacles on its path to becoming the de facto database for AI workloads, including competition from both relational databases and emerging AI-specific databases.

While I’m more optimistic than the street on MongoDB’s growth potential and, more importantly, its ability to expand FCF margins, I struggle to underwrite anything beyond a 25x FCF multiple (excluding SBC) for 2028. MDB's ability to command a FCF multiple over 25x will depend on how well it can fend off competition, especially as new database technologies tailored to AI workloads emerge. Traditional relational database providers like Oracle and Microsoft SQL are entrenched in enterprise markets, and hyperscalers offer their own database solutions that could compete more aggressively against MDB. Additionally, the rise of AI-specific databases or open-source alternatives could limit their ability to expand its addressable market as much as bullish scenarios predicts. A more crowded and competitive landscape could compress margins and slow growth, potentially lowering the achievable multiple. The price level where MongoDB starts to become attractive for me would be below $200 per share, which is where the stock traded in late 2022 and early 2023. At that level, the risk/reward balance would be much more favorable.