Pull to refresh
Logo
MongoDB buys Firebolt to run analytics inside its database

MongoDB buys Firebolt to run analytics inside its database

Money Moves

The database company is betting AI agents need sub-second answers on live data, not a warehouse a pipeline away.

Overview

Updated 23 minutes ago

MongoDB said on November 17, 2026 that it is buying Firebolt, an analytics database company, and will run its query engine next to MongoDB's operational database, synchronized automatically. Customers would get fast analytical queries on live data without building a pipeline into a separate warehouse. MongoDB calls the thing it removes the data-movement tax.

The pitch is aimed at AI agents. An agent rerouting a shipment or flagging a payment needs an answer in under a second, on data seconds old, thousands of times at once. A warehouse fed by an overnight pipeline cannot do that, and MongoDB has not said whether its engine avoids keeping a second copy of the data.

Why it matters

If analytics runs inside MongoDB's own database, the copy-and-sync pipeline most companies maintain for live dashboards becomes optional.

Questions about this story

Free account needed to ask — your question is kept and asked for you right after sign-up. Answers are public.

No questions yet — be the first to ask.

Key Indicators

Undisclosed
Purchase price
MongoDB gave no terms in its announcement or its securities filing.
$1.4B
Firebolt's prior private valuation
Calcalist reports the acquisition price is estimated to be significantly lower.
75%
Fortune 100 companies using MongoDB
MongoDB cites this figure as the reliability bar it plans to bring to analytics.
1 trillion+
Rows in Similarweb's Firebolt deployment
The customer-facing analytics workload runs at petabyte scale with sub-second queries.

Voices

Curated perspectives — historical figures and your fellow readers.

Sign up to generate historical perspectives on this story.

People Involved

Organizations Involved

Timeline

2022 November 2026

5 events Latest: 3 days ago
Tap a bar to jump to that date
  1. Price gap draws attention

    Latest Analysis

    Coverage notes Firebolt last raised at a $1.4 billion valuation while the purchase price is estimated to be significantly lower.

  2. Architecture question left open

    Scrutiny

    The Register asks MongoDB how data is stored and moved, noting automatic sync does not explain whether a second copy exists.

  3. MongoDB announces Firebolt acquisition

    Acquisition

    MongoDB says the Firebolt engine will run alongside its database, synchronized automatically, with no customer-managed pipeline. Terms were not disclosed.

  4. Firebolt confirms the deal

    Statement

    Firebolt says it is joining MongoDB and that integration work will take the coming months.

  5. MongoDB tries analytics beside the database

    Product

    MongoDB ships columnstore indexing so developers can run analytical queries without moving data to another system.

Scenarios

1

MongoDB turns on real-time analytics for customer data by mid-2027

Likely Resolves by Q2 2027

Discussed by: MongoDB's own announcement, which says customers will be able to switch it on in the coming months and has opened a waitlist

MongoDB says the Firebolt engine will run beside the database and stay in sync automatically, with no pipeline for customers to build. If that schedule holds, the first production users would appear in the first half of 2027.

2

MongoDB documents a separate data copy under the analytics engine

Possible Resolves by Q1 2027

Discussed by: The Register, which asked MongoDB to explain how the engine stores and moves data

Automatic synchronization says nothing about whether queries hit the original records or a copy kept in columnar form. If MongoDB later documents a separate store, the pitch narrows from no data movement to no pipeline you have to manage.

3

Firebolt's standalone product is retired and users pushed to MongoDB

Unlikely Resolves by End of 2028

Discussed by: No one publicly; MongoDB has said existing Firebolt customers keep their current service and support

MongoDB committed to unchanged service for Firebolt's existing customers. Setting an end-of-life date for the standalone product would reverse that commitment and shift the customer base onto MongoDB's platform.

4

Databricks ships one-copy transactions and analytics first

Possible Resolves by Q2 2027

Discussed by: The Register, which notes Databricks, SingleStore, Oracle and SAP all chase the same combination

Databricks said in 2026 that Reyden and Lakebase deliver transactions and analytics with one copy of the data, and SingleStore already sells a single system for both. A general-availability release from either would show MongoDB is following, not leading.

Historical Context

3 moments from history that rhyme with this story — and how they unfolded.

2010

SAP HANA (2010)

SAP shipped an in-memory database that held transactional and analytical data in the same column store, promising customers they could drop separate analytical systems. It became the core of SAP's product line.

Then

SAP's installed base adopted HANA as it migrated off older databases.

Now

Most large enterprises still kept separate warehouses for reporting and data science. One system for both workloads stayed a promise more than a practice.

Why this matters now

The claim that transactions and analytics belong in one place is 16 years old. The hard part has never been the idea; it is doing both without slowing either down.

2024

Databricks acquires Tabular (2024)

Databricks bought Tabular, the company founded by the creators of Apache Iceberg, the open table format used to manage large datasets on object storage. The deal consolidated control over a format many vendors depend on.

Then

Databricks positioned itself as the neutral home for open table formats.

Now

Data platform vendors kept buying their way into adjacent layers rather than building each one.

Why this matters now

MongoDB is making the same bet from the other direction: buying an analytics engine instead of building one, to keep customers inside its platform.

2022

MongoDB columnstore indexing (2022)

MongoDB introduced columnstore indexes so developers could run analytical queries against application data without changing the document structure or copying data elsewhere.

Then

It gave MongoDB a way to answer analytical queries inside the database.

Now

It did not stop customers from running separate warehouses for heavy analytical work.

Why this matters now

Firebolt is MongoDB's second attempt at the same problem, and this time it is buying a proven engine rather than adding an index type.

Sources

(8)