Steven Deobald, CEO
Håkan Råberg, CTO

Proposition

Endatabas is a

SQL Document Database

with Full History
 

"Never write your own database."

...so why did we?

Problem

A landscape of splintered and scattered data tools, compounding costs, and employee time wasted

It Can't Last

"But industry insiders don't expect this period of free-flowing cash, sky-high valuations and a litany of vendors to last.

Instead, they expect a wave of consolidation within the next decade"

$5000

Confluent Cloud

Event streaming, as a source of truth.

$2000

DynamoDB

1TB of storage, transactional workloads, and CDC

$1000

Aurora

1TB of storage, reserved 1x writer, 3x read replicas, backup

$12,000
/mo

$4000

Redshift

Standard configuration, 1 node, 24 months of storage

small
team
costs

Customer Pain

Every company wastes
money, time, and people
on data products.

In the 2020s, many data architectures feel like a 5-car garage.

Solution

Simplicity: One Database

Architecture

• Immutable data
• Time travel
• GDPR-compliance

Full History

• Dynamic columnar storage
• Separated storage / compute
• Adaptive indexes

SQL Document DB

UX / DX

A ground-up database allows the opportunity to build a familiar SQL built on SQLite — with power tools

-> INSERT INTO users (name) VALUES ("Tom")
-> INSERT INTO users {name: "Joseph"}





-> UPDATE users SET role = 'admin'
   WHERE id = '99bede7'



-> SELECT * FROM sales 
   FOR SYSTEM_TIME AS OF 2007-12-07T00:00:00



-> ERASE FROM accounts WHERE id = '4c69a89'



-> SELECT * FROM customers
   WHERE addresses..city = 'Stockholm'



-> SELECT AVG(price) FROM black_friday_sales

Immutable documents

...plain old SQL or convenience syntax

Non-destructive updates

("the database is the log")

Time travel

Native GDPR-compliance

Dynamic SQL

Columnar computation

Telemetry-based Debugging

Example GTM Use Cases

Reified Event Streams

OLTP Append-Only Time-Travel

Exploratory Analytics Queries

Warehousing Disparate Schemas

2023

After 14 months of development, Endatabas has a fully-functional core:

github.com/endatabas/endb

Benchmarks:

• SQL Logic Test
• SQL Acid Test (81%)
• TPC-H
• TPC-C (wip)

Downsizing: Many people would be happier with a Volvo 240 wagon.

Market

TAM = $80B+

$38b

2017

16% CAGR

- Gartner

$80b

2022

"unprecedented growth"
- Gartner

$262b

2030 - DBMS Market

Projected

DBMS
market growth

$472b

2030 - Cloud Storage Market

23.4% CAGR

- Fortune Business Insights

Customers want:

• a simpler product
• replace NoSQL stores with a SQL-native store
• integrate legacy schemas into one store

Everything in one place:

• immutable data in the large
• infinite storage
• HTAP

• start with small customers ($12,000/mo) who want to consolidate products
• dynamic SQL enables small teams in large corporations

• grow into medium-sized ($50,000/mo) who want HTAP and adaptive indexes

• leverage ground-up build to compete with Oracle, AWS before 2030

• "One Database" means users (developers, data scientists, etc.) stay in the product longer

• SaaS helps us grow the product in new ways:
  AI for adaptive indexes benefits from more data

• we sell the dream of Aurora, but actually help customers manage costs instead of becoming a new burden

Growth

MOA

Position

The time is right.

Start small and accessible.

Grow "One Database".

Win against largest players.

Sell Simplicity.

Our
Competitors

• Aurora, Neon (Postgres)
• MongoDB, Neo4j, Cosmos
• InfluxDB, Materialize
• Firebase, Fauna
• Snowflake, SingleStore

Not Our
Competitors

• Spanner, Cockroach
• ClickHouse
• Elastic
• VoltDB
• dbt

Competitors

• Aurora, Neon (Postgres)
• MongoDB, Neo4j, Cosmos
• InfluxDB, Materialize
• Firebase, Fauna
• Snowflake, SingleStore

Defensibility

• not just "another Postgres"
• SQL-native (not ad-hoc QL)
• time in the DB, not 2nd tool
• SQL+HTTP (not ad-hoc APIs)
• open source

A ground-up database started now wins in 2030.

• AGPL-3.0 - "Ultra Restrictive Copyleft"
• deep tech understanding
• deep market understanding

• dynamic SQL (early sales)
• HTAP (growth)

Why Endatabas Wins

Pricing

SaaS first

• simple pricing
• bundle compute + object storage into unit
• easier to start
• easier to grow

On-prem second

• $100,000/yr minimum
• support cannot be a consulting burden

f/oss

• analytics over valuable Endb data

• time-travel a major use case
• build apps on Endb

on-prem

• integration DB
• "cheap glue"
• consolidate schemas
• generate reports

on-prem

• reifying events
  (ingest Kafka, CDC)
• creating a timeline
• become a platform

SaaS

SaaS

• ad-hoc usage
• exploratory queries
• low barrier to entry
• dev tool, nice UX

Customer Adoption Journey

(Example)

Investment

Fund the Post-Postgres database today

$3m to $80b

2024

• find $3m USD seed
• hire 1-3 engineers
• pay for cloud infra

Exit Strategies

• Large vendor M&A: Oracle, Microsoft, IBM
• Service vendor M&A: Amazon, Google
• IPO as early Third Wave Commercial Open Source

2023

• build out HTAP to claim OLAP marketshare

• endb in all 4 major clouds
• hire sales team

2024

• hire 2nd engineer
• polish client UX
• adaptive indexes
• build DBaaS
• acquire initial customers

2025

• hire 3rd engineer
• hire DevRel
• attend conferences
• harden core
• endb multi-node

2026

2027

• bootstrap
• test our willpower
• build fully-functional core product

Investment
Timeline

Team

6 years of R&D

20 years of discovery

6 Years R&D

Decades of distributed systems development,
years of database research,
and three predecessors:
 

• Bank risk data lake
  (2017)
• XTDB 1.x Datalog Engine
  (2018 - 2020)
• XTDB 2.x SQL Engine
  (2021 - 2022)
• Endatabas
  (2023 - )

Founding Team

Steven Deobald

CEO

Håkan Råberg

CTO

Appendix

Technology

• Rust <=> Common Lisp
• SBCL Dynamic Runtime
• Apache Arrow
• HTTP + WebSocket APIs

• SQL Logic Test (SQLite)
• SQL Acid Test
• TPC-H, TPC-C

Future
Possibilities

• Vector Indexes
• Full-Text Search
• SQL:2023 (PGQ)
• Bitemporal Indexes

Questions?