Edelbyte
by Edelbyte
One engine for search, analytics, vector, graph and geospatial queries, built for AI agents that query company data constantly
Edelbyte is a data platform for enterprises whose AI agents, applications and analysts all need fast queries over very large datasets. Built by the former Microsoft team behind the Kusto engine (Azure Data Explorer), it combines search, analytics, vector, graph and geospatial workloads in one system so data teams can retire several specialized databases.
Edelbyte is a search and analytics data platform founded in November 2025 by four former Microsoft engineers who built Kusto, the engine behind Azure Data Explorer, Microsoft Sentinel, Microsoft Defender and Azure Monitor: CEO Avner Aharoni, CTO Evgeney Ryzhyk (who invented the Kusto engine and its query language), VP of Engineering Alexander Sloutsky and VP of Platform Engineering Yoni Leibowitz. The company came out of stealth on November 11, 2026 with about $70 million: a $25 million seed led by Sequoia and a $45 million Series A led by Quiet Capital, at a reported $500 million valuation, according to Calcalist. Its argument is that AI agents will send far more queries at a company's data than people ever did, so the engine indexes all data to answer ad hoc questions without anyone predicting access patterns in advance. One system handles text search, BI-style analytics, vector, graph and geospatial queries, includes a built-in knowledge base that gives agents context, and can federate queries to data where it already sits (Apache Iceberg, Parquet, S3, Kafka, PostgreSQL, MySQL). The company claims sub-100-millisecond queries on multi-petabyte datasets and deployments holding hundreds of petabytes; these are vendor claims with no public benchmark yet. It runs in public clouds, sovereign clouds or on premises, managed or self-hosted, and connects to Claude, Codex, Grafana, Tableau and Jupyter. Edelbyte says it already runs petabyte-scale systems for Fortune 500 customers in information security, automotive manufacturing and financial services, though none are named. It positions itself against Snowflake, Databricks and ClickHouse, with about 40 employees across Israel, Europe and the US.
A head of data platform or chief data officer running separate search, log analytics, vector and warehouse systems who expects agent query volume to multiply and wants one engine that can absorb it.
Interactive-speed queries over full-history, petabyte-scale data for people, applications and AI agents from a single governed engine.
At a Glance
- Category
- Data & Analytics
- Pricing
- Usage-based
- Target Market
- Chief Data Officers, CTOs, Data Platform Engineers, Security Operations Teams
- Deployment
- Multi-cloud, Self-hosted, Hybrid
- Founded
- 2025
- Headquarters
- Tel Aviv, Israel
- Team Size
- 11-50
- Customers
- Several unnamed Fortune 500 customers (company-reported, November 2026)
Key Features
- ✓Unified multi-model engine
Runs text search, analytics, vector, graph and geospatial queries in one system, so teams stop syncing data across several specialized databases.
- ✓Index-everything design
Indexes all ingested data and automates part of index tuning, so agents and analysts can ask ad hoc questions nobody planned for.
- ✓Query federation
Queries data where it already lives in Iceberg, Parquet, S3, Kafka, PostgreSQL or MySQL instead of forcing a full load first.
- ✓Hot cache plus long-term storage
Keeps recent data in a hot cache for millisecond queries and full history at object-storage cost, still queryable on the same engine.
- ✓Agent and BI connectivity
Connects to Claude, Codex, Grafana, Tableau and Jupyter so agents, dashboards and notebooks query one governed copy of the data.
- ✓Flexible deployment
Runs in public clouds, sovereign clouds, self-hosted environments or on premises, which matters for regulated and data-residency-bound workloads.
Capabilities
Use Cases
- •Security analytics
A security team queries complete authentication and transaction logs in real time to detect suspicious logins without sampling or aging out history.
- •Agent-facing data layer
An enterprise lets AI agents run thousands of concurrent queries against governed data, absorbing much higher query volume without a matching budget increase.
- •Customer-facing live metrics
A product team serves live figures inside a customer-facing application straight from the analytics engine, with millisecond response times claimed by the vendor.
- •Device telemetry analysis
An automotive or industrial manufacturer analyzes streaming device telemetry at petabyte scale and correlates it with search and geospatial queries in one place.
- •Observability consolidation
A platform team queries complete logs and traces from one engine and visualizes them in Grafana instead of running separate log and metrics stores.
Ideal For
Best For
- ✓Security and observability teams querying complete logs and traces at petabyte scale
- ✓Real-time detection of suspicious logins and transactions over streaming data
- ✓Giving AI agents governed, low-latency access to the same enterprise data analysts use
- ✓Consolidating separate search, vector, graph and analytics databases into one platform
- ✓Regulated or sovereign deployments that need the engine self-hosted or on premises
Not Ideal For
- ✗Small teams with gigabyte-scale data, where a managed warehouse or Postgres is simpler and the petabyte architecture brings no benefit
- ✗Buyers who need named reference customers, analyst coverage or independent benchmarks before shortlisting, since none were public at launch in November 2026
- ✗Organizations that want a mature partner ecosystem and large hiring pool today, which Snowflake and Databricks already have
Deployment
Market & Ratings
Several unnamed Fortune 500 customers (company-reported, November 2026)
Market Analysis
Pros
- ✓Founding team built and rebuilt Kusto, which gives the performance claims more credibility than a typical first-year database
- ✓Consolidating search, analytics and vector workloads can cut the number of systems a data team operates
- ✓Deployment options cover sovereign clouds and on premises, which many cloud-only warehouses do not offer
- ✓Usage-based pricing with no commitment lowers the cost of a pilot
Cons
- ✗No customers are named, and the 10x-queries-at-the-same-budget and sub-100ms claims are unverified vendor statements
- ✗No G2, PeerSpot, Hacker News or Reddit feedback existed as of November 17, 2026, so there is no practitioner signal on failure modes
- ✗Unit prices are not published, so total cost against Snowflake or ClickHouse cannot be compared without a quote
- ✗A roughly 40-person company is a concentration risk for a core data platform compared with established vendors
Pricing
Pay-as-you-go
Contact for pricing
- ✓No commitment
- ✓Query and ingestion compute billed per vCore-hour
- ✓Hot cache and long-term storage billed per GB-month
- ✓Networking and I/O charged at cloud cost
Edelbyte describes pay-as-you-go pricing with no commitment, metering compute per vCore-hour and hot cache and long-term storage per GB-month, with networking and I/O passed through at cloud cost. Unit rates were not published in the launch coverage reviewed on November 17, 2026, so expect a sales conversation for any real estimate.
Security & Compliance
Sources
This page was written from 7 sources, 6 on domains other than edelbyte.io.
- 1.sequoiacap.com — partnering with edelbyte data infrastructure that keeps up w
- 2.calcalistech.com — rk9dvkfrmx
- 3.siliconangle.com — siliconangle.com
- 4.unite.ai — unite.ai
- 5.aiweekly.co — edelbyte exits stealth with 70m seed plus series a from sequ
- 6.siliconreport.com — edelbyte emerges from stealth with 70m from sequoia quiet ca
- 7.edelbyte.io — edelbyte.iovendor
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
Revefi
AI FinOps and observability agents that cut LLM, Snowflake, Databricks and BigQuery spend
Corvic AI
Agentic data engine that turns PDFs, images, sensor logs and tables into cited answers and pipelines
Snorkel AI
Expert-built training data, RL environments and evaluation benchmarks for frontier labs and enterprises building specialized AI agents
Starburst Enterprise Intelligence Platform
Federated Trino-based platform and AIDA assistant that let AI agents query governed enterprise data in place, without moving it