MongoDB (MDB.US) Q2 conference call: Atlas is growing steadily, Enterprise Advanced unexpectedly becomes the “second engine”, and the adoption rate of AI products has risen significantly

The Zhitong Finance App learned that in the second quarter of the 2027 fiscal year ending August 2026, MongoDB (MDB.US) handed over an impressive report card. The company's total revenue reached US$772 million, a strong year-on-year increase of 30%. The core driving force was the accelerated growth of its cloud database platform Atlas and the significant acceleration of Enterprise Advanced self-management solutions. Based on continued optimism about enterprise-level demand, a steady increase in Atlas consumption, and early adoption of AI-related workloads, MongoDB simultaneously raised its annual performance guidelines.

CJ Desai, the company's president and CEO, said this quarter was the highest quarterly revenue growth rate the company has recorded since fiscal year 2024. During this period, the total number of MongoDB customers increased to 70,600, with a net increase of 2,900 customers in a single quarter, setting a record. At the same time, the company's non-GAAP operating margin reached 24%, and profitability continued to increase.

Atlas is growing steadily, and enterprise-level demand is strong

As the “ballast stone” of performance growth, the Atlas platform's revenue maintained a year-on-year growth rate of about 29% for the fifth consecutive quarter. Chief Financial Officer Mike Berry revealed that the growth rate exceeded the company's guidelines by about 300 basis points, and Atlas's revenue growth in a single quarter reached a record $127 million year over year. Growth was mainly driven by North American enterprise customers, particularly large customers with annual recurring revenue (ARR) of more than $100,000. The company's overall net ARR expansion rate increased to 122% from 119 percent last year and 121% in the previous quarter, showing continued deep exploration of the value of existing customers.

“Looking ahead, we expect large enterprise customers, particularly the US market, to continue to contribute very significant growth,” Berry emphasized during the Q&A session. He added that although AI-related demand is currently relatively small, it has shown encouraging growth.

At the end of the quarter, MongoDB had nearly 3,000 ARR customers exceeding $100,000, a year-on-year increase of 17%. Among the Atlas platform's customer base, 48% use two or more platform features (such as vector search and text search), up from 42% in the same period last year, indicating that the breadth of product portfolio adoption is increasing.

The AI product line has accumulated strength, and the adoption rate has increased significantly

Management highlighted the adoption of Atlas Vector Search and Voyage AI embedding and reordering models as strong evidence of increased AI workload activity. Berry notes that Voyage's customer numbers have nearly doubled for the second consecutive quarter, and that the vast majority of these new customers had no previous partnership with MongoDB.

Desai said that programming agents, including Anthropic's Claude and OpenAI's Codex, have become Voyage's main source of traffic. Additionally, the company has launched a managed model context protocol (MCP) server, which enables developers and AI agents to seamlessly connect to MongoDB through tools such as Claude Code, Codex, Grok Build, Cursor, and Devin.

In terms of customer stories, the Financial Times uses Atlas Vector Search and the Voyage model to provide its readers with AI-driven content discovery services, processing more than 100,000 queries per day. Desai said the publisher not only reduced search costs, but also shortened the manual index monitoring process, which previously took weeks to complete in one day. Fireflies, an AI workplace assistant platform, has chosen MongoDB since its inception and currently runs more than 40 microservices on it. Legal tech company EVE uses MongoDB's embedding and reordering capabilities to efficiently retrieve evidence from case files for its AI products.

Desai further pointed out that currently the most attractive AI application scenarios are mostly aimed at production-level, customer-interactive applications rather than small internal auxiliary tools. He specifically mentioned semantic search and wealth management consultant chatbots in financial services, as well as internal knowledge search applications for employees.

Enterprise Advanced becomes the new growth engine

The revenue of Enterprise Advanced and other businesses increased 36% year-on-year in the current quarter. When standardized according to contract terms, the ARR growth rate was about 11%. Berry said this was the strongest quarter for the business segment in the past three years. The company said that this success stems from the broad needs of customers in the financial services, public sector and technology industries, rather than a single high-value transaction. During this period, MongoDB also launched search and vector search capabilities for Enterprise Advanced, enabling self-managed customers to use search functionality for their AI workloads in a controlled environment.

Desai clarified that Enterprise Advanced's rapid growth was not at the expense of Atlas. He described the customer's deployment decision as “choosing both, not choosing one of the two,” and many organizations will use both platforms simultaneously due to hybrid cloud deployment, compliance requirements, and operational resilience.

On the case side, a major US bank has deployed Enterprise Advanced in more than 100 production applications and expanded it for generative AI and semantic search; the Nationwide U.K. (Nationwide U.K.) also uses Enterprise Advanced and Atlas to support its high-speed applications, processing more than 24 million application logins per week.

Berry revealed that in the current outlook, Atlas accounts for about 74% of total revenue, and this ratio is expected to continue to rise. However, he also emphasized that as an increasingly steady growth driver, Enterprise Advanced's contribution will exceed previous expectations.

Profit quality improved, year-round guidance raised

In terms of financial performance, MongoDB's non-GAAP operating profit for the quarter was US$186 million, and the operating margin reached 24%, far higher than 15% in the same period last year. Non-GAAP net profit was $163 million, or $1.90 per diluted share, compared to $87 million, or $1.00 per share for the same period last year. The company has achieved GAAP earnings per share for the third consecutive quarter. Total cash and short-term investments at the end of the period were US$2.4 billion, and operating cash flow and free cash flow were US$142 million and US$138 million, respectively.

For the third quarter of fiscal year 2027, MongoDB expects revenue of between $756 million and $761 million, up 20% to 21% year over year; non-GAAP operating profit of $152 million to $156 million; and non-GAAP earnings per share of $1.61 million.

Looking at the full year, the company raised its revenue guidance for fiscal year 2027 to $2.99 billion to $3.03 billion, corresponding to an increase of 21% to 23%. Among them, Atlas's annual revenue growth forecast was raised to about 27%, up 300 basis points from the previous forecast; the growth rate of Enterprise Advanced and other businesses was raised to about 11%, while previously it was only a mid-single digit increase. At the same time, the company raised its annual operating margin guidelines. It is expected to expand by about 250 basis points over the same period last year, and will continue to invest in AI, product research and development, market expansion, and developer ecosystem construction.

Conference content

Operator:

Hello, and welcome to MongoDB's Q2 FY2027 results conference call. Now I'm leaving the conference to Jess Lubert, VP of Investor Relations. Please get started.

Jess Lubert

Vice President of Investor Relations at MongoDB

Thanks, operator. Good afternoon everyone, and thank you for participating in our conference call today to review MongoDB's financial results for the second quarter of fiscal year 2027 — the relevant data was revealed in a press release issued after today's close. I am attending the conference today with CJ Desai, President and CEO of MongoDB, and Mike Berry, Chief Financial Officer.

In this conference, we will present forward-looking statements, including statements relating to the market and future growth opportunities, new business development expectations, Atlas consumption growth trends, the impact of EA and other business and multi-year license revenue, AI long-term opportunities, financial guidance and basic assumptions (including profitability and operating margin expectations), and our investment and growth opportunities in the AI field.

The above statement is affected by various risks and uncertainties, including factors such as operating performance and financial conditions, which may cause actual results to differ materially from our expectations. For a discussion of material risks and uncertainties that may affect actual results, please refer to the risk factors described in our quarterly report to the SEC for the period ending July 31, 2026 (Form 10-Q) submitted to the SEC on September 1, 2026. Any forward-looking statements made at this meeting represent our views as of today, and we are under no obligation to update them except as required by law.

Additionally, this conference call will discuss non-GAAP financial measures. Please refer to the attached tables in the results announcement posted on our Investor Relations website for reconciliation information between these metrics and the most directly comparable GAAP financial measures.

Well, now I would like to ask CJ to speak.

CJ Desai

President and CEO of MongoDB

Thank you Jess, and thank you all for participating today. I'm excited to share with you our excellent second quarter results. Total revenue reached $772 million, up 30% year over year, the highest quarterly growth rate since FY2024. Atlas's revenue continued to grow approximately 29% year-on-year for the fifth consecutive quarter, mainly due to contributions from large enterprise customers and continued accumulation of AI momentum. EA and other businesses performed well this quarter, growing 36% year over year, due to widespread demand from our ability to “run anywhere”. Driven by strong revenue growth, our non-GAAP operating margin reached 24%. The total number of customers reached 70,600 at the end of the quarter, and the net number of new customers added during the period reached a record of 2,900.

The number of Voyage customers nearly doubled month-on-month, and the growth rate of Atlas vector search adoption continued to surpass the company's other business segments, fully demonstrating our early strong momentum in AI workloads. Our core business remains strong, and the “run anywhere” advantage was a key differentiator in both Atlas and EA's growth this quarter. Businesses in financial services, healthcare, technology, etc. are running their most mission-critical workloads on MongoDB, and we're gaining more workloads every quarter. At the same time, these companies and AI native companies are increasingly choosing our platform to handle AI workloads, as evidenced by the use of Atlas vector search and Voyage AI embedding models. This quarter, I and my team continued to have in-depth discussions with the customer's C-level executives to discuss how our data platform can meet their most pressing core business and AI modernization needs.

The performance in the second quarter is the reason I firmly believe that we are emerging as a real-time intelligent data platform for modern applications in the multi-cloud and AI era. First, I'd like to talk about observations on the enterprise side. For customers already running large data assets on MongoDB, building an agent on top of this is a natural extension, because what the agent actually needs is real-time operational data rather than an old copy in the data warehouse. The search, vector search, and embedding functions are all natively built-in rather than plugged-in, so agents don't need to connect to multiple isolated systems; they only need to connect to a unified platform. This trend is reflected in a variety of use cases across industries, from internal knowledge retrieval, customer-facing chatbots and agents, to fraud identification and identity management workflows.

Although it's still in its early stages, we're seeing more and more of these workloads entering production environments. For example, the Financial Times uses our platform to support AI-driven content discovery to deliver interactive experiences to millions of readers at scale. With Vector Search and Voyage AI, the Financial Times has now unified its operational data and vector embedding on a single platform, built a hybrid full-text and semantic search solution, eliminating the complexity of synchronizing multiple systems and significantly reducing time to market. By using the high-precision Voyage-4 model for content indexing and using the cost-effective Voyage-4-Lite model to process more than 100,000 queries per day, the Financial Times minimized the impact on performance while drastically reducing search costs. Manual index monitoring, which used to take weeks, can now be completed within a day.

Next, let's talk about the momentum we're seeing in cutting-edge laboratories — these are both our customers and our partners. Leading labs are using Atlas to handle mission-critical workloads critical to their product launches. One of the labs used Atlas for inference and conversational workloads after migrating from PostgreSQL — previously PostgreSQL's performance delays and outages seriously affected the user experience. They migrated their conversation memory system to Atlas in just four weeks, and now reads 10 times faster than PostgreSQL. Additionally, these laboratories use Atlas for research workloads, storing experimental results, evaluation data, and training artifacts required for model development. These partnerships are still in their early stages, with varying degrees of participation from lab to lab, but we are excited about the progress. As a partner of these cutting-edge labs, we're helping developers building apps and agents on their platforms easily use Atlas.

Just recently, we launched a fully managed MCP server, making it easier for developers and agents to directly connect to MongoDB when using popular coding tools such as Claude Code, Codex, and Grok Build, as well as Cursor and Cognition's Devin. This is how we are deeply integrated into the AI supply chain and participate in the construction of next-generation applications. Commenting on our technical cooperation and recent integration with Claude, Paul Smith, Anthropic's Chief Commercial Officer, stated, “The best AI applications require powerful databases, which is why we've always recommended MongoDB Voyage to developers developing on Claude for embedding. Recently, demand from these developers has driven MongoDB to build a new hosted MCP server, which has been adopted extremely fast since launch, and now developers can explore, query, and manage their MongoDB data without leaving Claude.”

The final piece of the puzzle of AI opportunities is AI native companies. The data layer of these companies determines whether the product can support rapid scaling. Some companies choose us right from the start, while others start with other platforms (such as tip-driven development platforms) and migrate to us when they encounter expansion bottlenecks and actual usage comes in. This model is already reflected in the data — we added a record 2,900 new customers this quarter, many of which are AI-native companies. Fireflies is a Unicorn AI native startup dedicated to creating what it calls the “First AI Work Assistant” to help users uncover knowledge hidden in conversations. Fireflies serves more than 20 million users in more than 1 million organizations and has processed more than 7 billion minutes of meeting minutes. Fireflies chose Atlas from day one, valuing its flexible document model over a rigid relational model; today they run more than 40 microservices that use change streams to support real-time data pipelines for analysis and growth intelligence. This lean and scalable infrastructure seamlessly supports its ultra-rapid growth.

We're also seeing strong momentum on Voyage. Voyage is our embedding and reordering model, and continues to top independent rankings. In August, we introduced automated Voyage embedding to Atlas, enabling one-click vector search setups; released the Voyage Code 4 model designed specifically for code; and launched an upgraded reordering API — all of which have enabled Atlas to maintain its market-leading position in AI search accuracy. Voyage's appeal is evident on both sides of the market: some of our largest existing Atlas customers are starting to use Voyage for AI use cases, and most new Voyage customers are AI-native companies not previously associated with MongoDB. Eve is one of them, a unicorn AI native company that automates legal case acceptance, medical timeline sorting, and claim drafting for the plaintiff's law firm. Eve used the Atlas embedding and reordering API supported by Voyage AI's ReRank 2.5 to sift out the most relevant evidence from a large number of case documents. This not only improves EVE's RAG layer search quality, but also streamlines the infrastructure needed to build and iterate these AI experiences.

Switch to Enterprise Advanced (EA). EA's growth this quarter was broadly reflected in our installation base, particularly in the financial services, technology, and public sector sectors. Two models of customer use of EA are noteworthy, both of which illustrate the strategic importance of this business for us. The first category is AI applications in a regulated, self-managed environment. This quarter, we're bringing search and vector search capabilities to EA, bridging the gap between cloud services and self-managed experiences. Customers from various industries were immediately in demand, and they wanted to use an integrated solution to build AI in their own regulated, self-managed environment. A major US bank is a typical example. EA has become its standardized data platform, supporting more than 100 production applications such as payment, fraud detection, document processing, and customer and account services. This quarter, the bank expanded the same environment to generative AI and semantic search for employee advisors, chatbots, product search, and document intelligence. By unifying operational data, search, and vector retrieval in EA's self-managed environment, they keep sensitive customer and conversation data within their own regulated environment without sending it to multiple isolated systems. This provides them with a practical foundation for scaling AI across the industry based on the same platform—already running their most critical business.

The second category is hybrid deployment. My conversations with customers have increasingly addressed the need to run simultaneously across multi-cloud and self-managed environments. For customers using EA and Atlas at the same time, the two are “both” rather than “choose one of the two.” For example, a major cybersecurity company had significant deployments on both Atlas and EA, and both saw significant growth this quarter. The UK Nationwide (the world's largest building association) is also a good example. They now run rapidly growing speed-tier apps simultaneously on EA and Atlas, which can not only provide members with real-time access to account and transaction data through all digital channels, but also support more than 24 million app logins per week. The distribution of workloads across EA and Atlas gave Nationwide greater operational resilience, while meeting UK regulatory requirements and simplifying an otherwise more fragmented IT architecture. Nationwide is already using our platform to process faster payments, processing up to 3 million transactions worth £1.5 billion on peak days through a self-managed dual-cloud EA cluster. Bringing AI into a self-managed environment has opened up new requirements for us, and hybrid deployments often mean that a strong EA foundation brings new opportunities for Atlas partnerships with the same customers. At the same time, EA's profitability allows us to invest more heavily in R&D and marketing to further drive Atlas's growth and AI roadmap.

Finally, I have confidence in the leadership team driving innovation at Atlas and EA. Ben Cefalo is responsible for core products, and Pablo Stern is responsible for AI and emerging products. In terms of market expansion, John McMahon has quickly entered the role as the new Chief Revenue Officer, giving me full confidence in our ability to seize future opportunities. Before I finish, I'd like to remind everyone that we'll be hosting Investor Day in New York City on September 29th and a.local New York user event on September 30th. We look forward to seeing you all at that time.

Next, I'd like to ask Mike to speak.

Mike Berry

MongoDB Chief Financial Officer

Thanks CJ. Good afternoon everyone. I'll first review our results for the second quarter of fiscal year 2027 and then present our outlook for the third quarter and the rest of the fiscal year. As always, I'll be discussing GAAP and non-GAAP results at the same time.

As stated by CJ, we had another very strong quarter, with all indicators exceeding all guidance ranges. In view of this performance and the strong momentum across all business lines, we continued our excess performance in the second quarter and raised our guidance for the second half of fiscal year 2027, mainly due to the strong performance of Atlas.

Before going into details, I'd like to highlight a few key points for this quarter: First, total revenue growth increased to 30%, the first time since FY2024. Second, this is the fifth consecutive quarter that Atlas has maintained approximately 29% year-on-year growth. Third, EA and other businesses performed exceptionally well this quarter, growing 36% year over year, thanks to EA's growing strategic importance to many large customers and the early results of the EA search and vector search features we launched in the second quarter. Fourth, thanks to these trends, we have significantly exceeded expectations in terms of operating margin and earnings per share guidance, reflecting the advantages of our operating model.

Next, let's look at specific results. Total revenue for the second quarter was $772 million, up 30% year over year, compared to 24% in the same period last year. Looking at product segments, Atlas' revenue increased by about 29% year over year, exceeding the guideline by about 300 basis points. Consumer performance was strong, exceeding expectations for the third time in a row, in line with our guiding framework. This is the sixth consecutive quarter that Atlas achieved year-on-year growth in dollar amounts, with a record net increase of $127 million this quarter. The main drivers of growth this quarter were still North America and our largest customers, particularly those with annual recurring revenue (ARR) of over $100,000, which is in line with the upward market expansion momentum we have discussed in recent quarters. This continued strength is also reflected in the company's net ARR expansion rate, which rose to 122% this quarter, compared to 119% in the same period last year and 121% in the previous quarter. The month-on-month increase in net ARR expansion rate was driven by both Atlas and EA. We also continue to see positive momentum in the AI native community and various AI signals, including the adoption of vector search, the number of new Voyage customers, and the continued increase in clusters connected via MCP.

We have had a very strong performance in terms of EA and other revenue. Revenue increased about 30% year over year, making it the strongest quarter in three years. The EA search and vector search features we launched in the second quarter received early demand, and the additional search capabilities further enhanced our ability to support AI workloads. This growth is widespread, and is reflected in the joint drive of many transactions rather than a single major transaction, particularly in the financial services, public sector, and technology industries. Continued momentum highlights EA's strategic importance, and customers continue to expand their self-managed deployment scale to support both traditional and AI applications. EA and other ARRs (after standardized adjustments for the impact of contract terms) increased by about 11% year-on-year, achieving double-digit ARR growth for the third consecutive quarter.

Next, let's look at the income statement. Total non-GAAP gross margin was 75.9%, up about 210 basis points year over year; subscription gross margin was 78.3%, up about 70 basis points year over year. The increase in gross margin of subscriptions was mainly due to the increase in EA's share of revenue in the second quarter.

In terms of profitability, we are very happy to see that the second quarter was our third consecutive quarter of GAAP earnings per share, and the full-year guidance also includes our expectations for achieving GAAP earnings per share for the 2027 fiscal year. Non-GAAP operating revenue was $186 million, and operating margin was 24%, compared to 15% in the same period last year. We are satisfied with the continued performance of operating margins, which contributed significantly to the strong revenue performance this quarter. Non-GAAP net profit for the second quarter was $163 million, or $1.90 per share, based on 85.8 million fully diluted tradable shares. In comparison, net profit for the same period last year was US$87 million, or US$1.00 per share, based on 87.1 million fully diluted tradable shares.

Our remaining performance obligations (defined as contractual obligations over 12 months) ended the quarter at $1.52 billion, up 91% year over year, with a partial increase of 73% over the current period. New customer growth was strong this quarter, with approximately 2,900 new customers added month-on-month, bringing the total number of customers to 70,600, compared to 59,900 in the same period last year. Growth was mainly driven by Atlas. The number of Atlas customers reached 69,300 at the end of the second quarter compared to 58,500 in the same period last year. Within Atlas, the number of Voyage customers nearly doubled month-on-month for the second consecutive quarter, continuing to show positive signs of demand for our AI embedding capabilities. We are pleased with the momentum of new customers and are reminded that this indicator fluctuates from quarter to quarter.

At the end of the quarter, we had nearly 3,000 customers with at least $100,000 in annualized recurring revenue, up 17% year over year. This group's revenue growth is still strong. The growth rate exceeds the company's overall revenue growth rate, which is in line with our direction of expanding into the high-end market. Atlas's feature adoption also continues to be good. Of Atlas customers with annual recurring revenue of over $100,000, 48% are using two or more features of our platform, up from 42% in the same period last year, mainly driven by the adoption of vector and technology search.

In terms of balance sheet and cash flow, we held $2.4 billion in cash equivalents and short-term investments at the end of the second quarter. This quarter, we allocated $100 million to share buybacks and $59 million to settle taxes on employees' restricted stock units. Cash flow from operating activities was US$142 million, compared to US$72 million in the same period last year; free cash flow was US$138 million, compared to US$70 million in the same period last year. We are committed to driving meaningful and lasting cash flow. In the first half of fiscal year 2027, we have achieved cash flow from operating activities of US$344 million and free cash flow of US$335 million.

Now I'd like to share some of the assumptions driving the third-quarter outlook and add to our thoughts for the rest of the 2027 fiscal year. As I mentioned earlier, we continue to be happy with Atlas's strong and steady growth. Growth to date has been mainly driven by the continued strength of large enterprise customers, and we expect this trend to continue in the second half of fiscal year 2027. Based on this continued momentum, we expect Atlas to grow by around 26% in the third quarter and raise our full-year growth forecast to about 27%, up 300 basis points from the midpoint of previous guidance. The increase in overall revenue guidance for the second half of the year was mainly due to Atlas' strong performance. Atlas's strength is reflected in the sixth consecutive quarter of year-over-year revenue growth in dollar amounts, strong net ARR expansion rates, increasing multi-product penetration, and early signs of AI workload adoption. We've discussed over the past few quarters that as Atlas grows in size, it's more predictable and less sensitive to revenue fluctuations for individual customers or groups. This can be seen from the consistency of our results over the past three quarters — we have surpassed our initial guidance by approximately 200 to 300 basis points. We used the same guiding framework to develop our outlook for the third quarter, while recognizing that the third quarter was Atlas's most challenging comparison base for the whole year.

For EA and other businesses, given the strong performance in the first half of the year, including demand from the search and vector search functions launched on EA in the second quarter, we raised EA and other revenue expectations for the full year of fiscal year 2027 to a growth rate of about 11%, which is higher than the previous guidance for mid-single-digit growth. This is the first time in three years that EA and other businesses are expected to achieve double-digit growth for the full year. The guidance for the second half of the year is in line with what we shared last quarter. We continue to expect EA and other revenues to remain essentially flat in the second half of the year and achieve mid-single-digit growth in the third quarter. As multi-year contracts are inherently difficult to predict, we will continue to be cautious about EA's business guidelines. We are excited about EA's growth and recommend that you focus on full-year growth rather than single-quarter performance, as performance will naturally fluctuate from period to period.

In terms of profitability, the leveraging effect of the business model can be seen from the results of the first half of fiscal year 2027 — while continuing to invest in growth initiatives (particularly engineering and product innovation), it can still drive incremental profitability. We remain committed to the dual goals of driving revenue growth and increased profitability. We now expect the operating profit margin to expand by about 250 basis points for the 2027 fiscal year, which is 100 basis points higher than the previous upper limit of the range. As we achieve this expansion, we will continue to invest in key growth initiatives in product and marketing. Product investments are still focused on enhancing AI and core database capabilities, including related investments in EA. On the upcoming Investor Day, everyone will hear more about our new product innovations. In terms of marketing, we are investing more to accelerate the adoption of new product innovations and continue to focus on the fastest-growing opportunities in geographic regions and customer segments. We will also continue to invest in quota sales staff, marketing programs, and developer awareness building. In terms of cash flow, given the strong performance in the first half of the year, we now expect the full year free cash flow conversion rate to reach the upper end of our long-term target range of 80%-100%.

Now let's take a look at how this translates into guidance for the third quarter of fiscal year 2027. Once again, the increase in guidance for the second half of the year was mainly driven by the strength of Atlas. For the third quarter, we expect total revenue of $756 million to $761 million, an increase of 20% to 21% year over year. We expect non-GAAP operating revenue of $152 million to $156 million, and an operating margin of approximately 20.5% at the upper limit of the guideline. We expect non-GAAP net profit per share of $1.57 to 1.61 million, based on 87.1 million diluted tradable shares. For fiscal year 2027, we now expect total revenue of $2.99 billion to $3.03 billion, an increase of 21% to 23% for the full year. At the upper limit of the guideline, this will be the second consecutive year of accelerated growth in total revenue. We expect non-GAAP operating revenue of $616 million to $636 million, and an operating margin of approximately 21% at the upper limit of the guideline. Combined with 23% revenue growth and 21% operating margin, our goal is to achieve the “Rule of 44” performance based on the upper limit of the FY2027 guideline. We expect non-GAAP net profit per share of $6.39 to $6.58, based on 86.4 million diluted tradable shares.

Finally, I'd like to thank the entire MongoDB team for another excellent quarter of execution. We are satisfied with our results, confident in the sustainability of our growth, and will continue to focus on creating long-term value for our shareholders while investing responsibly in our business. Last but not least, we look forward to seeing you all on Investor Day later this month. You can find out more about registering for the live event or listening to the webcast on our Investor Relations site. Well, ask the operator to open the questioning session.

Q & A session

Operator:

Thank you. The first question came from Barclays Raimo Lenschow. Your line is connected.

Raimo Lenschow

Analyst, Barclays

OK, thank you. Congratulations on your excellent quarterly results. My question is about Atlas. If I watch -- or if I listen to your guiding reviews -- the strong performance was driven by Atlas. What factors are you considering? What gave you such confidence? Thank you.

Mike Berry

MongoDB Chief Financial Officer

Thanks for the question, Raimo. I'm Mike. As we mentioned, we feel really good about the Atlas business. What we are concerned about is that this is the fifth consecutive quarter of approximately 29% year-on-year growth, which is very stable. We raised our full-year guidance by 300 basis points from the previous one, which was also supported by a record net additional Atlas revenue of $127 million and an increase in net ARR expansion. Today, Atlas's annual operating rate is close to $2.3 billion. Looking ahead, we continue to expect growth from large enterprise customers, particularly in the US, to maintain a positive trend. We're already starting to see some of the benefits from AI, and although it's still small, we're excited about the momentum and expect consumption to continue to stay in line with the first half of the year.

Raimo Lenschow

Analyst, Barclays

Thank you.

Mike Berry

MongoDB Chief Financial Officer

Thank you.

Operator:

Thank you. The next question comes from Alex Zukin of Wolfe Research. Your line is connected.

Alex Zukin

Analyst, Wolfe Research

Hi guys. Thank you for taking the questions. I think CJ, if I look at business, it's clear that Atlas accelerated in the first half of the year, and subscription revenue growth also accelerated. However, it feels like Atlas may slow down slightly in the second quarter. The guidance for the rest of the year, and the fourth quarter in particular, suggests that Atlas will experience a fairly significant deceleration. I understand conservative principles, but if we're still in the early stages, as AI native companies and labs are building up their potential, what are some of these dynamics? Is there a situation where EA turned some of the deals into Atlas like what happened in the fourth quarter of last year? What's a dynamic we're probably not seeing?

CJ Desai

President and CEO of MongoDB

OK, Alex. Thank you. Let me answer that your question contains quite a few points. First, it's extremely encouraging to see Atlas now continuing to grow at 29%, as Mike outlined, and overall it's very exciting for us to execute, both on the enterprise side and within the AI native community. As you said, compared to the early March guidance, we saw an acceleration in the first half of the year — this is the first point. Second, I want to be very clear that EA (Self-Managed MongoDB) did not grow at the expense of Atlas. Atlas is actually still continuing to grow, Alex, we meet the needs of our customers where they are.

When I first joined and outlined in the first earnings call, the customer asked us that we wanted to run these large, heavy workloads on MongoDB EA. “CJ, we want to be ready for AI, so the team should build search and vector search on that.” This type of workload will run in our self-managed environment for a variety of reasons — whether it's data sovereignty or they don't want to move to the public cloud for any other reason. We did, and delivered the feature on June 30th. We've seen a very good response from our customer base. As Mike mentioned, our self-managed MongoDB growth was broad, not focused on a single customer, but across industries.

The first point is that I am very optimistic about Atlas consumption trends from the first half of the year to the second half of the year. The second point is the growth we've seen in self-managed EA — whether they're running on a new cloud, running in an on-premises data center, or sometimes some customers run the EA in a specific region of the public cloud — we feel this isn't at the expense of Atlas. In the few examples I've highlighted, we've actually seen that from an operational resilience perspective, some large bank or government customers say this is an advantage of the MongoDB data platform rather than choosing between Atlas and EA.

Regarding the guidelines, I'll let Mike add, but we've raised the Atlas guidelines for the whole year by 300 basis points. You know we started in March and are now growing 27%. We will always be cautious. As for the fourth quarter, it still depends on consumption trends. From our point of view, there's still some time left. We need to watch the performance in September and October — this will be the baseline. Then the fourth season vacation season arrives, which will indeed affect our spending. So we're trying to be cautious in our guidance, and I'm optimistic about what I'm seeing, both in terms of the Atlas core group and AI natives.

Mike, would you like to add to the Atlas guide?

Mike Berry

MongoDB Chief Financial Officer

Yes. Thank you. Very well answered, CJ. I just wanted to emphasize what he said — our guiding philosophy, Alex Zukin, has not changed in how we guide the rest of the year. We are always cautious about timeframes longer than a quarter, and this is reflected in the guidelines. I'd also like to respond to your comments about the fourth quarter and make it very clear on the conference call: there were no big bundle deals this quarter. This quarter did not show the same dynamics as the fourth quarter of last year.

Operator:

Thank you. Please wait for the next question. The next question comes from Goldman Sachs's Matt Martino. Your line is connected.

Matt Martino

Analyst, Goldman Sachs

OK. Thanks for taking my questions. CJ, this is the second consecutive quarter you've highlighted the strong momentum in the number of Voyage customers, and I think you mentioned an interesting observation in preparation for your presentation — quite a few Voyage customers are brand new to MongoDB. This seems like a great funnel top to win over AI's native hyper-growth workloads. How would you describe the results of turning these customers into sales on a wider range of platforms so far? Thank you.

CJ Desai

President and CEO of MongoDB

OK, Matt. I'll talk about it from two perspectives. The first perspective is that we're really excited about the number of new customers we've received through Voyage. You're absolutely right, and that's exactly why I made it clear in my statement — many of them really aren't current MongoDB customers. This is a fact. As you know, the acquisition was completed in February 2025, so we're only 18 months away. The Voyage team is making sure we have the best embedding model for things like benchmarking external data. Most importantly, some customers came in as Voyage customers and then became Atlas customers. It's still too early, because we're only just starting to make sure we can cross-sell and upsell — whatever words you want to use.

But I think this is a huge opportunity for Atlas's long-term growth, and we're getting these customers through Voyage. This is almost always the case, and when I look at the customer names we get — whether in the San Francisco Bay Area, large enterprises, London, Tel Aviv, or Seattle — they tend to be driven by AI workloads. As the team analyzed Voyage's recommendation sources, as you might imagine, most of the recommendations came from coding agents—Claude first and then Codex, which drove most of Voyage's referral traffic. So we have multiple benefits: coding agents love Voyage, they recommend us, and we get a new customer base as a result.

The second point is that this is at the top of the funnel, and as you said, we'll be cross-selling and upselling. The Alliance team developed plans for Atlas customers. Third, these are almost all AI workloads. So overall, it's still early, but it's very encouraging.

Operator:

Thank you. The next question comes from UBS's Karl Keirstead. Your line is connected.

Karl Keirstead

Analyst, UBS

OK, thank you. Maybe I'd like to ask CJ directly about this question. CJ, you mentioned that MongoDB is already seeing some early momentum in terms of AI workloads. As we all try to monitor the timing and scale of the upcoming AI pull on MongoDB, I was wondering if you could elaborate on what types of AI use cases or workloads you've found to be driving MongoDB the most? I'm particularly interested in one of Mike's mid-season reviews — he mentioned customer-facing enterprise workloads. I know you mentioned customer support use cases in preparation for your statement, but maybe you can go into more detail about the specific use cases that have the strongest pull effect so we can focus on these aspects. Thank you.

CJ Desai

President and CEO of MongoDB

Of course you can. Based on my observations from many conversations with customers, Karl, I'll look at the business side first. The “enterprise” here refers to the world's top 2000 or Fortune 500 or Fortune 100. Looking at these businesses, MongoDB is almost always used as a database platform for customer workloads — whether it's insurance claims, health insurance policies, credit card transactions, or providing fast data to end users. MongoDB has always performed well when handling large-scale, customer-facing workloads. What I'm seeing so far is, let's say you're a wealth manager at a bank, and have lots of knowledge base articles that need to be vectorized, use our embedded model, and then use it as a chatbot for wealth management advisors to help them communicate with customers in real time. This is a very specific example, where a large bank is using MongoDB. There are other examples, such as because of the large number of documents, knowledge base use cases for employees, that employees can use this to search — because search is now fully integrated into operational data and document loading, and embedding makes vectorization even better. This is another large enterprise example where we see use cases. But my clear feeling is that it's almost always: “We want to use MongoDB when the agents we create for customer-facing activities need to scale massively” — no matter what those customer-facing activities are.

What we haven't seen yet are use cases like “hey, I created a little assistant for a few hundred employees” getting early traction. We haven't seen MongoDB being used for this kind of scenario because customers would say “this is too small” — of course MongoDB provides scale and all the other features. This is the first point. Karl, the second point I want to make is that when you look at AI native companies — I'm going to include cutting-edge labs — but in the example I'm sharing, whether it's the EU or Fireflies.ai, they're all intelligence entities in production environments. When you look at these smart bodies in production — we've mentioned ElevenLabs and others in the past — these are the millions of agents that perform customer-facing tasks in production, they'll say, “We want to use MongoDB in terms of scale, performance, and the ability to 'run anywhere'” — that's what they're for. These are the two main directions I see: enterprise side, smart devices are put into production and run on Atlas in the right scenario; on the EA side, I mentioned in my speech that, unquestionably, EA's growth is being driven across industries — including the tech industry — as these regulated industries work to prepare their operating data for AI. Customers will say, “Hey, I'm building AI agents for my technology platform” — no matter what technology platform uses MongoDB, or the tech company itself — we've seen this kind of growth.

That's my general summary of what I've observed. If anything else needs to be added, I'll give Mike the floor.

Mike Berry

MongoDB Chief Financial Officer

Nothing to add. Very well answered. Thank you.

Operator:

Thank you. Please wait for the next question. The next question comes from Morgan Stanley's Sanjit Singh. Your line is connected.

Sanjit Singh

Analyst, Morgan Stanley

OK, thanks for taking the question. CJ, I want to focus on Enterprise Advanced because EA's growth curve has improved markedly during your tenure, while Atlas's growth — as you mentioned — continues to remain at a very attractive level of 29%. Some of the information we've heard from customers is that if MongoDB wants these customers to eventually move to Atlas, it's important to advance EA's capabilities — including search and vector search, and possibly Voyage AI. What's your opinion on this: First, what is the timeline for these customers to adopt the new AI features? What might the timing be to eventually “upgrade” or “migrate” to Atlas if you guys want to? This is the first part of the EA problem. The second part is, which new customer groups are using EA? You mentioned big companies, financial institutions — it's no surprise. Have you seen an opportunity — and I think you've also entered the new cloud space through EA — is there a possibility that AI native companies will also start using EA due to data sovereignty or other reasons? These are my two questions about EA. Thank you very much.

CJ Desai

President and CEO of MongoDB

OK. Sanjit, I'd like to raise my perspective a little bit first—we're a highly customer-driven company. The reason we're investing in the EA roadmap we've outlined — and I'm happy to see it working — is because customers told us that even early on, you have to invest in EA. If EA can prepare for AI through search, vector search, etc., they'll also ask, “Hey, can you also use Voyage AI in a self-managed environment?” It's entirely customer-driven, and we satisfy our customers where they are. This is the first point.

Second, as Mike shared, EA achieved three consecutive quarters of double-digit growth, which made me optimistic that we now have two growth engines — Atlas and EA. As I said, the two are not mutually exclusive, because this is very important. Sometimes customers say, “I need to do this for operational resilience.” Sometimes customers say, “I don't see the need to migrate this workload to Atlas, but given what you've done — particularly unifying search and vector search — this is a new workload we'd like to try on Atlas.” So our momentum at EA is also driving additional use cases for banks or public sector organizations on Atlas. Specifically, in Q2, we had a public sector customer decide to expand our use of self-managed MongoDB (as an EA). Meanwhile, we're currently working with them because they've seen some of the benefits of the MongoDB codebase, and they said, “CJ, if you guys manage and provide security patches and all the other services...” They're currently running an Atlas pilot project in their government cloud. From my point of view, it's very encouraging that MongoDB now has two growth engines, Atlas and EA.

Regarding the timeline: First, when we launched the search and vector search features — which were driven by AI requirements — we were charging customers an additional fee for this set of features available in EA. This is the first point. Second, with regard to time to achieve value, Sanjit, I think value is realized pretty fast — not months, but weeks — based on how we release these features to our customers. It's just that they self-manage, and we manage it in Atlas. Like the Financial Times case I shared, they used Atlas vector search and embedding to achieve value in weeks, not months or years, to prepare AI for applications such as search. That's my overall opinion.

Finally, I'd like to say that specifically in the banking and healthcare sector, what I'm seeing at the same time is: “Hey, CJ, we'll use Atlas, but we may also fail over to EA because of the operational resilience you provide — it's definitely world-class, and it's a huge advantage for us.” So in summary, I think this is MongoDB's enduring growth driver, driven by customer demand, which is “we want to run everywhere” — sometimes self-managed, sometimes managed by MongoDB.

Operator:

Thank you. The next question comes from Wells Fargo's Ryan MacWilliams. Your line is connected.

Ryan MacWilliams

Analyst, Wells Fargo

Hey, thanks for taking the question. There are two issues here. First question for CJ: Are you seeing customers approach you ahead of schedule renewals and renew at higher rates compared to a year ago? Are they making replenishment payments earlier, or are consumer trends improving more strongly? Second question for Mike: I know your guiding philosophy hasn't changed, but given the short history of quarterly Atlas guidance, some people have questions about the implied Q4 Atlas guidance. Can you help us understand the input factors for this guide and how investors should think? Thank you.

CJ Desai

President and CEO of MongoDB

OK. First, as Mike outlined, our NRR is very strong and high — this is true for both Atlas and EA. Regarding retention rates and my observations, even when customers may want to optimize dynamics such as workloads with our customer success team, the trend is very healthy and improving, which fully proves our 8.0 release last year — customers are very satisfied with the price/performance ratio and the way their consumption is growing. Now, some large customers are telling me that our consumption is growing faster than they expected and asked, “CJ, can we revisit our contracts if we continue to grow at this rate?” But this situation isn't common — sporadic, in the highest number of digits, and not widespread. This is encouraging and means we are not facing resistance to renegotiate contracts or commitments as consumption increases. That's my answer. Regarding guidance, I'll repeat what I said in response to Alex before — I feel very good about the Atlas business, its sustainability, and the innovation we drive. Of course, we won't be guiding the third quarter based on Mike's framework; we will always be cautious about the fourth quarter. But we've raised our guidelines — which is why the Atlas guidance for the whole year is 27%. mike?

Mike Berry

MongoDB Chief Financial Officer

Yes. Thanks CJ. Ryan, on this point, the methodology and philosophy of the guide remained the same throughout the year, and we want to stick to that — that is, when we guide the current quarter (or the next one), we always want to stay at — hey, you should refer to the 200-300 basis point range we provided. Hoping for better consumer performance, we can finally reach the upper end of this range. And for further quarters, Ryan, we'll always be cautious. It's a consumer business. I know it doesn't seem far away — just one more quarter — but we want to be cautious. I hope we perform well in the third quarter so that we can improve our guidance as we enter the fourth quarter.

Operator:

Thank you. The next question comes from Kirk Materne of Evercore ISI. Your line is connected.

Kirk Materne

Analyst, Evercore ISI

OK. Thank you so much for taking the questions. CJ, my question is about Voyage and Vector's add-on rates. I'm curious, when you get new customers with these products, do they experiment and then scale quickly, or what? Obviously, getting new customers through them is great, but scaling them up and making ARR more meaningful from an overall company perspective is what you want to achieve. I'm a little curious how fast these products are going from departmental pilots to being considered strategic or company-wide like Atlas or EA. Thank you.

CJ Desai

President and CEO of MongoDB

OK, of course. I'll talk about both aspects. Atlas is entirely consumption-driven, and you know that very well. When I see some big clients — like a few Fortune 100 customers — they've done all the testing and then say, “OK, we're searching in another isolated system” or “we're trying to do vector searches from an early startup that provides vector functionality.” This large bank told me that, based on their tests, they think vectors should be fully integrated into the operational data layer, and MongoDB sees this as a huge advantage. The value is realized there in a few weeks, then we will have a dedicated search node, etc., to drive consumption. Even a large media company — one of our biggest vector search customers — had agents try to do semantic queries and find out: “OK, if this is the operational data layer, then it works.” We've also seen a very good response from the AI native community that vectors as part of the database have received a very good response. One example I shared last quarter — ElevenLabs, which continues to scale well on MongoDB — saw vector embedding as a huge advantage. We're doing the same thing now in EA.

On the embedding side, we're making it easier to use — as I shared in my presentation, we've implemented automatic embedding in Atlas and how it works in the cloud. Currently, in all customer conversations, what I see is that there is still a low level of awareness that Voyage actually comes from MongoDB. A customer told me, “Oh, we love Voyage, we're using Voyage.” I said, “Did you know it's a MongoDB product?” They answered, “Oh, we don't know.” OK, then we should take a look at Atlas now, since you guys have Atlas automatically embedded. Situations vary. Search vector search, I don't think it will take too long to realize value because there are fewer moving parts involved — that is, the integration of multiple systems — which is why it works.

Operator:

Thank you. Ladies and gentlemen, due to time constraints, we will accept two more questions. The next question comes from Citi's Tyler Radke. Your line is connected.

Tyler Radke

Analyst, Citi

OK. Thank you. CJ, you mentioned some inference workloads in AI labs. Can you elaborate - are these new additions this quarter? How do you see the scale of these workloads compared to other large workloads at traditional companies? Mike, with regard to EA, clearly overperformed this quarter. I think you guys released a new feature in July from a GA perspective. What makes you believe there won't be more upside in the second half of the year, considering that most of the increase in guidance for the second half comes from Atlas? Thank you.

CJ Desai

President and CEO of MongoDB

OK. Tyler, we're very specific about our reviews because we want to be highly transparent to everyone. For one of the labs, they started an inference workload late in the second half of last year and ran it on MongoDB, using it as a memory layer. For that lab, our team observed how well Atlas performed on that specific inference workload, and then they said, “Wow, compared to PostgreSQL, Atlas performs really well in terms of reading and writing” -- which is exactly what they used to be. Then they recently -- in the second quarter, I think -- moved a few more inference workloads to Atlas for some of the other products they created. So we started an inference workload around November and December of last year, then they migrated several other inference workloads to other products around June and July. They told me bluntly — this is the tech team — Atlas took all of their uptime and performance worries away: “We didn't even have to think about it. Now when we create a new product, we want to run reasoning on it.” So what I'm saying is, we started with an inference, then some other workloads, then got some additional ones in the second quarter.

Mike Berry

MongoDB Chief Financial Officer

Tyler, here's Mike to answer your EA questions. Great question, thank you. Over the past three quarters, we've seen ARR growth — as CJ mentioned — in double digits. We've actually raised our full-year guidance from mid-single digits to 11%. You know, just like us, the difficulty here is estimating multi-year contracts. For the benefit of all of us, we are always careful to ensure that we are not overly optimistic until those deals land. If the history of the past four quarters is indicative, hopefully some of them will actually appear on multi-year contracts or on a larger scale than we expected. So of course, we want to make sure — especially for EAs — that they are careful in their guidelines. But as CJ mentioned, we're very happy with the progress there. We believe it can and will be an enduring engine of growth. Hope we can do better than what we have guided.

Operator:

Thank you. The last question came from Bank of America's Koji Ikeda. Your line is connected.

Koji Ikeda

Analyst, Bank of America

OK. Hey guys. Thanks for taking the time to ask questions. I want to ask about EA, and actually about Atlas and the business as a whole. Obviously, in preparation for the presentation and in your responses to all the questions, AI definitely sounds like it's becoming a driver for the overall business, and EA sounds really good too. CJ, I think you mentioned EA not at the expense of Atlas, but how should we think about Atlas and EA, and how might the final Atlas' revenue ratio of total revenue change over the next three to five years? Thank you.

Mike Berry

MongoDB Chief Financial Officer

Hey, Koji. I'm Mike. It's clear we'll have more discussions when we publish the guidelines next year. We will be hosting a special session on finance on Investor Day. If you look at this year's full-year guidance, Atlas is 27%, and EA is now 11%. If Atlas now accounts for about 74%, then it should definitely continue to increase its share. But we do expect EA to be a more enduring engine of growth. To some extent, it should continue to grow — probably not as fast as we thought before, because as you mentioned, the driving force of AI is in both Atlas and EA, and we're very happy that this is a “combination” rather than a “pick of two.” We expect Atlas to continue to account for a percentage of total revenue, but of course EA also contributed much more than we thought at the beginning of the year.

CJ Desai

President and CEO of MongoDB

Yes. Koji, from a technical perspective, “run anywhere,” operational resilience, hybrid multicloud — these are all different terms our customers use from us. What we saw in particular — someone asked about the new cloud and other aspects — yes. What happens is that when someone wants to run in a new cloud because they might face capacity issues in the public cloud, they say, “Hey, can we run EA in that new cloud?” This reflects our ability to “run anywhere” and drives demand. We are also providing a database as a service in some other new clouds, which also falls under EA's revenue line. That's why we're making it very clear that EA isn't at the expense of Atlas. Seeing this widespread trend — not just one, two, or three or four customers — is very encouraging for these two to become enduring engines of growth.

Operator:

Thank you. Ladies and gentlemen, I would now like to return the phone to management to conclude.

CJ Desai

President and CEO of MongoDB

Thank you very much, operator. In summary, we had strong results in the second quarter, with Atlas, EA, and AI workloads all showing broad growth. Notably, the breadth of demand — cutting-edge laboratories, global banks, the public sector, rapidly expanding startups, some expanding existing deployments, others new customers. We're seeing AI workloads being implemented in MongoDB in all areas. This is why we raised our outlook for the second half of the year, and why we are confident that we will continue to invest while continuing to expand operating margins. MongoDB is emerging as the go-to real-time intelligent data platform, and I've never been more confident about where we are positioned among our customers. Thank you all so much.

Operator:

Ladies and gentlemen, that concludes today's conference call. Thank you all for participating. You can now disconnect.