# Starburst Data

Latest posts from Starburst Data

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Your AI Agent Is Only Seeing Half The Answer

DevFeed: [Your AI Agent Is Only Seeing Half The Answer](<https://devfeed.tech/articles/your-ai-agent-is-only-seeing-half-the-answer-145830.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/your-ai-agent-is-only-seeing-half-the-answer>)

Published: 2026-11-19T18:00:05Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [ai-data-engine](<https://devfeed.tech/tags/ai-data-engine.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [business](<https://devfeed.tech/tags/business.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [connectors](<https://devfeed.tech/tags/connectors.md>), [console](<https://devfeed.tech/tags/console.md>), [content](<https://devfeed.tech/tags/content.md>), [context](<https://devfeed.tech/tags/context.md>), [contracts](<https://devfeed.tech/tags/contracts.md>), [control](<https://devfeed.tech/tags/control.md>)

### AI overview

Starburst AIDA and NetApp AIDE are presented as a way to answer questions that require both database values and document context, such as matching contract renewal dates and spend with notice clauses. AIDA plans the question, sends structured queries through Starburst's federated SQL engine and document searches through AIDE's MCP server, then joins the results and cites source versions. Starburst is described as querying structured and semi-structured data in place across 50+ connectors, with row- and column-level controls, masking, and auditing. AIDE makes files on NetApp storage searchable, preserves file permissions, and exposes content through MCP; its Metadata Services catalog reads metadata without copying files or using GPUs. The stated tradeoff is that this approach avoids a separate external index, but its document capabilities depend on files being accessible through AIDE and NetApp storage interfaces. The excerpt recommends starting with one cross-source question and a joint architecture workshop.

### Source excerpt

Most teams building AI agents hit the same problem. Everything isn't in the same place. ...

## To Own Your Intelligence Layer You Need an Enterprise Intelligence Platform

DevFeed: [To Own Your Intelligence Layer You Need an Enterprise Intelligence Platform](<https://devfeed.tech/articles/to-own-your-intelligence-layer-you-need-an-enterprise-intelligence-platform-141976.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/to-own-your-intelligence-layer-you-need-an-enterprise-intelligence-platform>)

Published: 2026-11-18T16:22:15Z

Content type: opinion

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [data-governance](<https://devfeed.tech/topics/data-governance.md>)

Tags: [access-controls](<https://devfeed.tech/tags/access-controls.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [context](<https://devfeed.tech/tags/context.md>), [data](<https://devfeed.tech/tags/data.md>), [data-context](<https://devfeed.tech/tags/data-context.md>), [data-intelligence](<https://devfeed.tech/tags/data-intelligence.md>), [data-platform](<https://devfeed.tech/tags/data-platform.md>), [enterprise-ai](<https://devfeed.tech/tags/enterprise-ai.md>), [enterprise-intelligence](<https://devfeed.tech/tags/enterprise-intelligence.md>)

### AI overview

The excerpt argues that enterprise agents need defensible answers across distributed systems, which requires access control across the whole task, consistent interpretation of business terms, and a trace of what the agent read and did. It says centralized platforms cover only their own systems, standalone catalogs can drift from query execution, and copying data into one place creates another estate to secure and govern. Starburst presents its Enterprise Intelligence Platform as a layer above existing platforms that supplies governed data and certified context at query time, with visibility into access and answer logic. Its products include Context Studio and AIDA for context and agent workflows, Mission Control and LakeOps for operations, Galaxy BYOC for deployment in a customer's AWS account, and Torin for GPU acceleration; the excerpt reports up to 7x faster performance in Starburst benchmarks.

### Source excerpt

We are entering a new phase of enterprise AI. In every industry, organizations are moving ...

## The Agent Can Guess. The Query Engine Can't.

DevFeed: [The Agent Can Guess. The Query Engine Can't.](<https://devfeed.tech/articles/the-agent-can-guess-the-query-engine-can-t-141975.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/gpu-acceleration-gpu-native-engine-torin-the-agent-can-guess-the-query-engine-cant>)

Published: 2026-11-18T16:05:22Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Query (disambiguation)](<https://devfeed.tech/topics/query.md>), [AI response variability](<https://devfeed.tech/topics/ai-response-variability.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [added](<https://devfeed.tech/tags/added.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [all](<https://devfeed.tech/tags/all.md>), [also](<https://devfeed.tech/tags/also.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analyze](<https://devfeed.tech/tags/analyze.md>), [announcement](<https://devfeed.tech/tags/announcement.md>), [apache](<https://devfeed.tech/tags/apache.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [application](<https://devfeed.tech/tags/application.md>), [can](<https://devfeed.tech/tags/can.md>), [engine](<https://devfeed.tech/tags/engine.md>), [query](<https://devfeed.tech/tags/query.md>), [query-engine](<https://devfeed.tech/tags/query-engine.md>), [question](<https://devfeed.tech/tags/question.md>)

### AI overview

Torin is Starburst's GPU-native SQL query engine, designed to accelerate analytical work while returning the same rows as CPU execution. Trino continues to parse, optimize, and schedule queries; Torin runs supported operators on NVIDIA GPUs, while unsupported work stays on the CPU in the same query. It preserves CPU-compatible behavior by analyzing expressions for failure risks--for example, avoiding division by zero when a filter condition guards a division--and choosing safe evaluation strategies. `EXPLAIN ANALYZE` shows where query work ran. Torin is best suited to large scans, joins, aggregations, AI data preparation, and text or pattern filters over Parquet tables in Apache Iceberg, Hive, or Delta Lake. It requires a supported NVIDIA GPU with at least 24 GB memory; 80 GB or more is recommended. `DISTINCT` in `SELECT`, user-defined functions, complex types, non-Parquet scans, and inter-worker exchanges remain on CPU. Performance gains vary from none to an order of magnitude; small or lightweight queries may not keep a GPU busy. Torin is in limited general availability for Starburst Enterprise.

### Source excerpt

Ask an AI agent the same question twice and you may get two different answers. ...

## Enterprise AI Needs Access to Data and Context. Your Platform Can Vouch for Only Its Own.

DevFeed: [Enterprise AI Needs Access to Data and Context. Your Platform Can Vouch for Only Its Own.](<https://devfeed.tech/articles/enterprise-ai-needs-access-to-data-and-context-your-platform-can-vouch-for-only-its-own-136453.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/enterprise-ai-needs-access-to-data-and-context-your-platform-can-vouch-for-only-its-own>)

Published: 2026-11-16T16:22:43Z

Content type: opinion

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [data-governance](<https://devfeed.tech/topics/data-governance.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [context](<https://devfeed.tech/tags/context.md>), [data](<https://devfeed.tech/tags/data.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [enterprise-ai](<https://devfeed.tech/tags/enterprise-ai.md>), [platform](<https://devfeed.tech/tags/platform.md>)

### AI overview

The article argues that enterprise agents need shared, approved business definitions and task-wide auditability to operate with less human checking. Definitions for terms such as "customer" or "flagged" can differ across CRM, billing and warehouse platforms; rebuilding them in each vendor's ontology risks drift, and each platform can vouch only for its own step. The proposed approach is to harvest existing definitions from catalogs, semantic layers and transformation code, have an accountable person review and approve them once, then provide the certified definition, relevant joins and rules at query time across systems. This is presented as a way to reduce repeated definition work and help analysts answer auditors' questions about which definition was used, what the agent accessed and in what order. The excerpt gives no implementation details or independent evidence for the stated survey figures.

### Source excerpt

Right now, every vendor wants to sell you an ontology. Each one will say it ...

## How Are Companies Handling AI Governance in Practice?

DevFeed: [How Are Companies Handling AI Governance in Practice?](<https://devfeed.tech/articles/how-are-companies-handling-ai-governance-in-practice-133240.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/how-are-companies-handling-ai-governance-in-practice>)

Published: 2026-11-13T05:00:24Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [Sovereign AI](<https://devfeed.tech/topics/sovereign-ai.md>)

Tags: [access-controls](<https://devfeed.tech/tags/access-controls.md>), [act](<https://devfeed.tech/tags/act.md>), [added](<https://devfeed.tech/tags/added.md>), [adoption](<https://devfeed.tech/tags/adoption.md>), [after](<https://devfeed.tech/tags/after.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-application](<https://devfeed.tech/tags/ai-application.md>), [ai-governance](<https://devfeed.tech/tags/ai-governance.md>), [ai-projects](<https://devfeed.tech/tags/ai-projects.md>), [ai-prompts](<https://devfeed.tech/tags/ai-prompts.md>), [ai-ready-data](<https://devfeed.tech/tags/ai-ready-data.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [easy](<https://devfeed.tech/tags/easy.md>), [governance](<https://devfeed.tech/tags/governance.md>), [querying](<https://devfeed.tech/tags/querying.md>), [write](<https://devfeed.tech/tags/write.md>)

### AI overview

The article's central recommendation is to enforce AI governance at the data access layer, where each agent query can be checked against role, purpose, and permission and recorded for audit. It describes four practices: federated access to govern data across cloud and on-premises systems without first centralizing it; shared business context and definitions for agents and analysts; read-only defaults, RBAC/ABAC, and detailed audit trails; and designing around sovereignty, residency, privacy, and risk-based regulatory requirements from the start. Guardrails add a separate behavioral layer: they complement row- and column-level access controls by limiting prompts, topics, or agent and data-product behavior. The text reports that structured business context raised one foundation model's enterprise-question accuracy from roughly 50% to over 90%, while providing no details on how that result was measured.

### Source excerpt

AI governance is easy to write down and hard to enforce once agents start querying ...

## Why AI Guardrails Matter

DevFeed: [Why AI Guardrails Matter](<https://devfeed.tech/articles/why-ai-guardrails-matter-133241.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/why-ai-guardrails-matter>)

Published: 2026-11-12T22:28:17Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [LLM security](<https://devfeed.tech/topics/llm-security.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [Data Management](<https://devfeed.tech/topics/data-management.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [access-controls](<https://devfeed.tech/tags/access-controls.md>), [actions](<https://devfeed.tech/tags/actions.md>), [admin](<https://devfeed.tech/tags/admin.md>), [administrator](<https://devfeed.tech/tags/administrator.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-accuracy](<https://devfeed.tech/tags/ai-accuracy.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-governance](<https://devfeed.tech/tags/ai-governance.md>), [ai-guardrails](<https://devfeed.tech/tags/ai-guardrails.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-stack](<https://devfeed.tech/tags/ai-stack.md>), [ai-systems](<https://devfeed.tech/tags/ai-systems.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [guardrails](<https://devfeed.tech/tags/guardrails.md>)

### AI overview

AI agents can act across systems and may follow malicious instructions, so traditional access controls alone do not address their behavior. The excerpt describes guardrails at the tool/API, retrieval, and output stages to keep agents in scope, resist manipulation, restrict sensitive topics, and make refusals visible. Guardrails depend on governed data products and auditable identity delegation; they complement rather than replace data governance. Starburst AIDA provides four controls: agent protection and data product protection are on by default; prompt limiting rejects overlong messages before they reach the model (default 50,000 characters); topic filtering is off until configured and blocks topics semantically. AIDA identifies the control behind a refusal and makes guardrail events searchable. The article presents conservative defaults as deployable and adjustable, while noting that controls need tuning for specific needs.

### Source excerpt

Your AI agent delivers significant value to your customers and employees. But what happens when ...

## Data Strategy vs AI Strategy: Are They the Same Thing?

DevFeed: [Data Strategy vs AI Strategy: Are They the Same Thing?](<https://devfeed.tech/articles/data-strategy-vs-ai-strategy-are-they-the-same-thing-130898.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/data-strategy-vs-ai-strategy-are-they-the-same-thing>)

Published: 2026-11-12T16:45:29Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Strategy](<https://devfeed.tech/topics/ai-strategy.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>)

Tags: [access-controls](<https://devfeed.tech/tags/access-controls.md>), [agent](<https://devfeed.tech/tags/agent.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-data-governance](<https://devfeed.tech/tags/ai-data-governance.md>), [ai-readiness](<https://devfeed.tech/tags/ai-readiness.md>), [ai-ready-data](<https://devfeed.tech/tags/ai-ready-data.md>), [ai-strategy](<https://devfeed.tech/tags/ai-strategy.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [analytics-workloads](<https://devfeed.tech/tags/analytics-workloads.md>), [applications](<https://devfeed.tech/tags/applications.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [audit](<https://devfeed.tech/tags/audit.md>), [business-context](<https://devfeed.tech/tags/business-context.md>), [collaboration](<https://devfeed.tech/tags/collaboration.md>), [compliance](<https://devfeed.tech/tags/compliance.md>), [data](<https://devfeed.tech/tags/data.md>), [data-strategy](<https://devfeed.tech/tags/data-strategy.md>), [strategy](<https://devfeed.tech/tags/strategy.md>)

### AI overview

An AI strategy extends the existing data strategy: it applies the same access, shared business context, and governance foundation to AI use cases, with higher standards for data quality and freshness because models or agents may act without a person checking first. The text argues that separate plans can create duplicate policies, parallel data copies, and conflicting metric definitions. It recommends federated access to data where it lives, reusable data products with consistent meanings, and governance controls that extend across analytics, applications, and AI. To align the strategies, audit existing data locations, access, and controls; map AI use cases to identify specific gaps; and designate one governance layer as the policy system of record.

### Source excerpt

When a board asks a CDO or CIO for an AI strategy, it's tempting to ...

## Will AI Replace Data Governance?

DevFeed: [Will AI Replace Data Governance?](<https://devfeed.tech/articles/will-ai-replace-data-governance-127083.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/will-ai-replace-data-governance>)

Published: 2026-11-10T00:26:49Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [data-governance](<https://devfeed.tech/topics/data-governance.md>), [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [shadow AI](<https://devfeed.tech/topics/shadow-ai.md>), [trust](<https://devfeed.tech/topics/trust.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [access-controls](<https://devfeed.tech/tags/access-controls.md>), [adoption](<https://devfeed.tech/tags/adoption.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-adoption](<https://devfeed.tech/tags/ai-adoption.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-and-data](<https://devfeed.tech/tags/ai-and-data.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [ai-projects](<https://devfeed.tech/tags/ai-projects.md>), [ai-ready-data](<https://devfeed.tech/tags/ai-ready-data.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [data](<https://devfeed.tech/tags/data.md>), [data-governance](<https://devfeed.tech/tags/data-governance.md>), [governance](<https://devfeed.tech/tags/governance.md>)

### AI overview

AI does not remove the need for data governance; it changes enforcement from mainly human review to automated policy applied at query time as agents access more sources. The article argues that ungoverned data can amplify inconsistent definitions, expose sensitive information, and produce unreliable AI answers. It identifies four building blocks: federated enforcement where data resides, fine-grained attribute-based access controls, governed Data Products with ownership and lineage, and audit logs of model and agent activity. Its checklist recommends extending controls to agent and MCP queries, automating metadata-driven policies, and auditing access. The article presents federated access paired with governance as an approach to moving AI projects toward production, citing Vectra AI and a NetApp partnership as examples.

### Source excerpt

What is the relationship between AI and data governance, and how is this connection shifting ...

## Why Your AI Agent Is at Risk of Not Accessing Your On-Premises Data

DevFeed: [Why Your AI Agent Is at Risk of Not Accessing Your On-Premises Data](<https://devfeed.tech/articles/why-your-ai-agent-is-at-risk-of-not-accessing-your-on-premises-data-122215.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/ai-agents-on-premises-data>)

Published: 2026-11-06T22:05:56Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [digital sovereignty](<https://devfeed.tech/topics/digital-sovereignty.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>), [data mobility](<https://devfeed.tech/topics/data-mobility.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [access-controls](<https://devfeed.tech/tags/access-controls.md>), [accessible](<https://devfeed.tech/tags/accessible.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-applications](<https://devfeed.tech/tags/ai-applications.md>), [ai-chatbot](<https://devfeed.tech/tags/ai-chatbot.md>), [ai-context](<https://devfeed.tech/tags/ai-context.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [ai-systems](<https://devfeed.tech/tags/ai-systems.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [air-gapped](<https://devfeed.tech/tags/air-gapped.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [context](<https://devfeed.tech/tags/context.md>), [data](<https://devfeed.tech/tags/data.md>), [risk](<https://devfeed.tech/tags/risk.md>)

### AI overview

The article argues that AI agents may fail or guess when they cannot access on-premises data and its business context, which may need to stay within a network or national boundary for security, regulatory, or sovereignty reasons. It presents two deployment paths: connect cloud agents to on-premises data through a secure private link, or run models and agents in-house in an air-gapped environment for more sensitive workloads. In either case, it proposes federated access through governed connectors, with role-based controls, rather than requiring all data to be centralized. A context layer built from curated, versioned data products is presented as a way to give agents meaningful, governed datasets. The article describes Starburst as supporting hybrid sources, on-premises Apache Iceberg tables, MCP access for agents, and natural-language access through AIDA, with controls, lineage, and audit logging through a shared access layer.

### Source excerpt

Your AI agent is only as valuable as the data and context it can reach. ...

## Why Keep a Data Warehouse If You Are Adopting a Data Lakehouse?

DevFeed: [Why Keep a Data Warehouse If You Are Adopting a Data Lakehouse?](<https://devfeed.tech/articles/why-keep-a-data-warehouse-if-you-are-adopting-a-data-lakehouse-121498.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/why-keep-a-data-warehouse-if-you-are-adopting-a-data-lakehouse>)

Published: 2026-11-05T22:52:02Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [data lake](<https://devfeed.tech/topics/data-lake.md>), [Amazon Redshift](<https://devfeed.tech/topics/amazon-redshift.md>), [data migrations](<https://devfeed.tech/topics/data-migrations.md>), [Data pipelines](<https://devfeed.tech/topics/data-pipelines.md>), [Open Data](<https://devfeed.tech/topics/open-data.md>), [etl](<https://devfeed.tech/topics/etl.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache](<https://devfeed.tech/tags/apache.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [apache-iceberg-tables](<https://devfeed.tech/tags/apache-iceberg-tables.md>), [batch](<https://devfeed.tech/tags/batch.md>), [bi](<https://devfeed.tech/tags/bi.md>), [bigquery](<https://devfeed.tech/tags/bigquery.md>), [business-intelligence](<https://devfeed.tech/tags/business-intelligence.md>), [cloud-data-warehouse](<https://devfeed.tech/tags/cloud-data-warehouse.md>), [concurrency](<https://devfeed.tech/tags/concurrency.md>), [concurrent](<https://devfeed.tech/tags/concurrent.md>), [cost](<https://devfeed.tech/tags/cost.md>), [costs](<https://devfeed.tech/tags/costs.md>), [data](<https://devfeed.tech/tags/data.md>), [data-lakehouse](<https://devfeed.tech/tags/data-lakehouse.md>), [data-warehouse](<https://devfeed.tech/tags/data-warehouse.md>), [lakehouse](<https://devfeed.tech/tags/lakehouse.md>), [warehouse](<https://devfeed.tech/tags/warehouse.md>)

### AI overview

The article recommends keeping a cloud data warehouse and an Iceberg-based lakehouse side by side, matching each workload to its needs and moving workloads gradually. Warehouses such as Snowflake, Redshift, and BigQuery remain suited to small, stable, governed data and high-concurrency BI, where managed performance may justify proprietary storage and indexing. Their costs can grow faster than useful query work due to idle or oversized clusters, egress, and duplicated data and pipelines. Open Iceberg tables add schema enforcement, ACID transactions, and time travel on object storage; they may better fit large or semi-structured data and shared access across teams or platforms. Iceberg's optimistic concurrency suits batch and merge-style writes, so frequent small-row writes need evaluation. The suggested low-risk approach is to query both platforms in place through a federated SQL layer, compare cost and performance on candidate workloads, then migrate incrementally based on evidence.

### Source excerpt

Adopting a data lakehouse doesn't mean retiring the cloud data warehouse on day one. Most ...

## The Lakehouse Isn't Enough. Agents Need an Intelligence Layer of Their Own.

DevFeed: [The Lakehouse Isn't Enough. Agents Need an Intelligence Layer of Their Own.](<https://devfeed.tech/articles/the-lakehouse-isn-t-enough-agents-need-an-intelligence-layer-of-their-own-109926.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/the-lakehouse-isnt-enough-agents-need-an-intelligence-layer-of-their-own>)

Published: 2026-10-30T18:27:05Z

Content type: opinion

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [ai observability](<https://devfeed.tech/topics/ai-observability.md>), [Data Management](<https://devfeed.tech/topics/data-management.md>), [AI for data automation](<https://devfeed.tech/topics/ai-for-data-automation.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [all](<https://devfeed.tech/tags/all.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [analyst](<https://devfeed.tech/tags/analyst.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-agents](<https://devfeed.tech/tags/autonomous-agents.md>), [autonomy](<https://devfeed.tech/tags/autonomy.md>), [cto](<https://devfeed.tech/tags/cto.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [intelligence](<https://devfeed.tech/tags/intelligence.md>), [isn-t](<https://devfeed.tech/tags/isn-t.md>), [lakehouse](<https://devfeed.tech/tags/lakehouse.md>), [layer](<https://devfeed.tech/tags/layer.md>), [martin](<https://devfeed.tech/tags/martin.md>), [need](<https://devfeed.tech/tags/need.md>)

### AI overview

The excerpt argues that autonomous agents need an intelligence layer above individual data platforms because a lakehouse alone cannot govern work that crosses systems. It frames defensibility around three needs: shared business meaning, access checks across the task, and a trace linking what the agent read and did. In its fraud example spanning PostgreSQL, Snowflake, and Salesforce, separate platform rules and logs leave gaps; the proposed layer would use a certified definition of "flagged," check access before a Salesforce read, and provide one trace across all three systems. The excerpt presents this as a way for analysts to stand behind agent answers, while acknowledging that accuracy alone is insufficient.

### Source excerpt

Our CTO, Martin Traverso, and SVP of Engineering, Jitender Aswani, recently argued that AI agents ...

## Not So Fast: Equality Deletes, Compaction, and Iceberg V4

DevFeed: [Not So Fast: Equality Deletes, Compaction, and Iceberg V4](<https://devfeed.tech/articles/not-so-fast-equality-deletes-compaction-and-iceberg-v4-106401.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/equality-deletes-compaction-and-iceberg-v4>)

Published: 2026-10-29T08:00:23Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>), [database-optimization](<https://devfeed.tech/topics/database-optimization.md>), [rocksdb](<https://devfeed.tech/topics/rocksdb.md>)

Tags: [compaction](<https://devfeed.tech/tags/compaction.md>), [iceberg](<https://devfeed.tech/tags/iceberg.md>)

### AI overview

Equality deletes let streaming CDC writers delete by key without locating the old row, and their meaning survives compaction that moves rows. Iceberg V4's October 22, 2026 draft prohibits new equality deletes but still allows reading existing ones; it remains under development. In the V3 paths reviewed, background conversion to deletion vectors (DVs) moves row-location lookup out of ingestion, but DVs tied to files can conflict with compaction. The V3 commit path does not automatically merge a concurrent DV into a prepared one: the converter must detect target changes, refresh its index, and recompute. Staggering jobs, shared locks, smaller rewrite groups, and partial progress can reduce conflicts, but may delay target-branch visibility or still fail under frequent updates. LakeOps reports batching equality-delete application to bound memory at the cost of extra local I/O. The excerpt argues that a replacement should preserve sustained concurrent progress between ingestion and maintenance; it presents rewrite repair as a design direction, not a current V3 capability.

### Source excerpt

Equality deletes are expensive to read, easy to accumulate, and awkward for row lineage and ...

## Is Data Federation Fast Enough for AI?

DevFeed: [Is Data Federation Fast Enough for AI?](<https://devfeed.tech/articles/is-data-federation-fast-enough-for-ai-102223.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/is-data-federation-fast-enough-for-ai>)

Published: 2026-10-27T08:00:39Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [data-federation](<https://devfeed.tech/topics/data-federation.md>), [presto](<https://devfeed.tech/topics/presto.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [High-Performance Computing](<https://devfeed.tech/topics/high-performance-computing.md>), [RAPIDS](<https://devfeed.tech/topics/rapids.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-performance](<https://devfeed.tech/tags/ai-performance.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [blackwell](<https://devfeed.tech/tags/blackwell.md>), [bloom-filter](<https://devfeed.tech/tags/bloom-filter.md>), [caching](<https://devfeed.tech/tags/caching.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [clusters](<https://devfeed.tech/tags/clusters.md>), [columnar](<https://devfeed.tech/tags/columnar.md>), [complex](<https://devfeed.tech/tags/complex.md>), [compute](<https://devfeed.tech/tags/compute.md>), [concurrency](<https://devfeed.tech/tags/concurrency.md>), [connect](<https://devfeed.tech/tags/connect.md>), [context](<https://devfeed.tech/tags/context.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [cpus](<https://devfeed.tech/tags/cpus.md>), [data](<https://devfeed.tech/tags/data.md>), [data-federation](<https://devfeed.tech/tags/data-federation.md>), [latency](<https://devfeed.tech/tags/latency.md>)

### AI overview

The article argues that modern federation can support near real time AI queries by reducing data movement and distributing work. Engines push filters, selected columns and some aggregations to source systems, use cost-based planning and dynamic filtering for cross-source joins, and process partitions in parallel. It recommends pairing federation for fresh operational data with Apache Iceberg for large historical scans and high-concurrency analytics. The text gives Starburst Torin as a GPU option: it runs existing Trino SQL over data in place and claims up to 7x faster performance than CPU baselines. Performance depends on workload and source systems; the article says controls such as concurrency limits, timeouts and read-only replicas help protect operational databases.

### Source excerpt

A common misperception about data federation is latency. The conventional wisdom states that querying data ...

## Torin Deep Dive: Lazy Evals in GPU-Native SQL Engine

DevFeed: [Torin Deep Dive: Lazy Evals in GPU-Native SQL Engine](<https://devfeed.tech/articles/torin-deep-dive-lazy-evals-in-gpu-native-sql-engine-95122.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/torin-starburst-nvidia-deep-dive-lazy-evals-in-gpu-native-sql-engine>)

Published: 2026-10-22T08:00:33Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [SQL](<https://devfeed.tech/topics/sql.md>), [GPU rendering](<https://devfeed.tech/topics/gpu-rendering.md>)

Tags: [cudf](<https://devfeed.tech/tags/cudf.md>), [engine](<https://devfeed.tech/tags/engine.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [lazy](<https://devfeed.tech/tags/lazy.md>), [sql](<https://devfeed.tech/tags/sql.md>), [starburst](<https://devfeed.tech/tags/starburst.md>)

### AI overview

Torin preserves SQL's short-circuit behavior for fallible expressions while keeping safe filters on its fast, column-at-a-time GPU path. It classifies expressions by how they can fail: eager evaluation for infallible expressions; NULL masking when NULL inputs prevent failure; and filter-and-scatter when masking cannot make evaluation safe, at the cost of copying active rows and restoring results. Torin routes lazy constructs such as AND, OR, COALESCE, IF and CASE through a shared IF mechanism, using LET to avoid recomputing values. It rewrites only where needed, so comparison-only filters remain eager. The text describes the design and tradeoffs but provides no performance measurements for the strategies.

### Source excerpt

Torin is the new GPU execution engine we built at Starburst. It runs SQL query ...

## AI Agents Are Outgrowing the Converged Data Platform

DevFeed: [AI Agents Are Outgrowing the Converged Data Platform](<https://devfeed.tech/articles/ai-agents-are-outgrowing-the-converged-data-platform-92731.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/nvidia-starburst-ai-agents-are-outgrowing-the-converged-data-platform>)

Published: 2026-10-21T11:30:59Z

Content type: opinion

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>), [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>)

Tags: [about](<https://devfeed.tech/tags/about.md>), [access-controls](<https://devfeed.tech/tags/access-controls.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [all](<https://devfeed.tech/tags/all.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [also](<https://devfeed.tech/tags/also.md>), [analysts](<https://devfeed.tech/tags/analysts.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [applications](<https://devfeed.tech/tags/applications.md>), [apply](<https://devfeed.tech/tags/apply.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [data](<https://devfeed.tech/tags/data.md>), [data-platform](<https://devfeed.tech/tags/data-platform.md>), [next](<https://devfeed.tech/tags/next.md>), [platform](<https://devfeed.tech/tags/platform.md>), [question](<https://devfeed.tech/tags/question.md>)

### AI overview

AI agents make governed access to distributed data and workload-appropriate query compute central to the data layer. The article argues for querying data where it lives, applying access controls, policies, and lineage to each query, and choosing CPU or GPU execution without changing the data architecture. Starburst Torin adds an NVIDIA GPU option alongside Starburst's CPU engine: supported SQL operations run on the GPU and unsupported ones continue on CPU, with existing SQL and applications retained. Starburst reports benchmark speedups on analytical workloads, while noting they do not measure full agent runs, results vary, and gains depend on the share of task time spent waiting on queries. Torin is limited general availability and runs on NVIDIA GPUs.

### Source excerpt

A person asks a question, reads the answer, and decides what to ask next. An ...

## Your Enterprise Data Wasn't Built for AI: Data Federation Can Help

DevFeed: [Your Enterprise Data Wasn't Built for AI: Data Federation Can Help](<https://devfeed.tech/articles/your-enterprise-data-wasn-t-built-for-ai-data-federation-can-help-90929.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/your-enterprise-data-wasnt-built-for-ai-data-federation-can-help>)

Published: 2026-10-20T08:00:16Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Integration Strategies](<https://devfeed.tech/topics/ai-integration-strategies.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [data-governance](<https://devfeed.tech/topics/data-governance.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [Large Language Model](<https://devfeed.tech/topics/llm.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>), [SQL](<https://devfeed.tech/topics/sql.md>), [Hallucination detection](<https://devfeed.tech/topics/hallucination-detection.md>), [presto](<https://devfeed.tech/topics/presto.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-data](<https://devfeed.tech/tags/ai-data.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [autonomous-ai-agents](<https://devfeed.tech/tags/autonomous-ai-agents.md>), [can](<https://devfeed.tech/tags/can.md>), [data](<https://devfeed.tech/tags/data.md>), [data-federation](<https://devfeed.tech/tags/data-federation.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [enterprise-data](<https://devfeed.tech/tags/enterprise-data.md>), [help](<https://devfeed.tech/tags/help.md>), [language](<https://devfeed.tech/tags/language.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [large](<https://devfeed.tech/tags/large.md>), [large-language-models](<https://devfeed.tech/tags/large-language-models.md>), [large-language-models-llms](<https://devfeed.tech/tags/large-language-models-llms.md>), [llms](<https://devfeed.tech/tags/llms.md>), [models](<https://devfeed.tech/tags/models.md>), [organizations](<https://devfeed.tech/tags/organizations.md>)

### AI overview

The article argues that data federation can give AI agents fresher, cross-system context while leaving data in its source systems. It describes legacy batch pipelines as risking stale results, schema misinterpretation, and security gaps. A federated engine queries sources in place, joins results, and applies centralized policies such as row filters and column masking. The article recommends exposing certified data products with explicit metric definitions and domain context, and enforcing read-only query rules server-side. It presents federation as complementary to warehouses and lakehouses: high-volume analytical workloads may fit open formats such as Apache Iceberg, while operational data can remain federated. The text promotes Starburst, built on Trino, as an implementation and claims query optimization can support sub-second interactive responses; it provides no independent performance evidence.

### Source excerpt

Large language models (LLMs) and autonomous AI agents represent a fundamental shift in how organizations ...

## What Is Data Access?

DevFeed: [What Is Data Access?](<https://devfeed.tech/articles/what-is-data-access-90088.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/what-is-data-access>)

Published: 2026-10-19T08:00:57Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Access Control](<https://devfeed.tech/topics/access-control.md>), [Authorization](<https://devfeed.tech/topics/authorization.md>), [Dynamic authorization rules](<https://devfeed.tech/topics/dynamic-authorization-rules.md>), [Security](<https://devfeed.tech/topics/security.md>), [identity and access management](<https://devfeed.tech/topics/identity-and-access-management.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [agent-identity](<https://devfeed.tech/tags/agent-identity.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agents](<https://devfeed.tech/tags/agents.md>), [aggregation](<https://devfeed.tech/tags/aggregation.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-models](<https://devfeed.tech/tags/ai-models.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [analyst](<https://devfeed.tech/tags/analyst.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [api](<https://devfeed.tech/tags/api.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [architectures](<https://devfeed.tech/tags/architectures.md>), [authentication](<https://devfeed.tech/tags/authentication.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [autonomous](<https://devfeed.tech/tags/autonomous.md>), [data](<https://devfeed.tech/tags/data.md>), [what-is](<https://devfeed.tech/tags/what-is.md>)

### AI overview

A data access layer gives users, analytics tools, and AI agents a governed way to query data across distributed sources. It centralizes identity checks, permissions, and query execution without requiring all records to be copied into one warehouse; federated engines can push filters and projections to source systems to reduce data transfer. The article contrasts RBAC, which assigns permissions by job role but can lead to role proliferation, with ABAC, which evaluates contextual attributes at query time. It also describes row filtering and column masking as runtime controls. For AI agents, it recommends certified data products with semantic metadata and strict server-side read-only enforcement: query parsing should reject writes and unauthorized statement types. Starburst, powered by Trino, is presented as one implementation for federated queries and shared security policies.

### Source excerpt

Data access is the data architectural capability that allows users, analytical tools, and AI agents ...

## Apache Iceberg Architecture Explained

DevFeed: [Apache Iceberg Architecture Explained](<https://devfeed.tech/articles/apache-iceberg-architecture-explained-88575.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/apache-iceberg-architecture-explained>)

Published: 2026-10-16T08:00:04Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>), [parquet](<https://devfeed.tech/topics/parquet.md>), [data lake](<https://devfeed.tech/topics/data-lake.md>), [table-partitioning](<https://devfeed.tech/topics/table-partitioning.md>), [Amazon S3](<https://devfeed.tech/topics/amazon-s3.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache](<https://devfeed.tech/tags/apache.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [apache-iceberg-architecture](<https://devfeed.tech/tags/apache-iceberg-architecture.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [format](<https://devfeed.tech/tags/format.md>), [high-performance](<https://devfeed.tech/tags/high-performance.md>), [iceberg](<https://devfeed.tech/tags/iceberg.md>), [object](<https://devfeed.tech/tags/object.md>), [object-storage](<https://devfeed.tech/tags/object-storage.md>), [open](<https://devfeed.tech/tags/open.md>), [performance](<https://devfeed.tech/tags/performance.md>), [storage](<https://devfeed.tech/tags/storage.md>), [table](<https://devfeed.tech/tags/table.md>), [table-format](<https://devfeed.tech/tags/table-format.md>)

### AI overview

Apache Iceberg tracks tables through a catalog pointer to immutable metadata and data files, rather than inferring table contents from directory layouts. Reads use snapshot and manifest statistics to prune files; writes create new files and metadata, then use an optimistic, atomic catalog pointer swap. This supports concurrent engine access, snapshots and time travel, while row level deletes and updates use delete files. The document distinguishes v2 position and equality deletes from v3 deletion vectors, which replace position deletes, and notes v3's VARIANT type and other additions. Engine and catalog compatibility must be checked before upgrading. Ongoing maintenance--compaction, snapshot expiration, orphan file removal and manifest optimization--helps control small file fragmentation, metadata growth and storage costs.

### Source excerpt

Apache Iceberg is an open table format designed for high-performance analytics running on object storage. ...

## How to Architect an On-Premises Data Lakehouse

DevFeed: [How to Architect an On-Premises Data Lakehouse](<https://devfeed.tech/articles/how-to-architect-an-on-premises-data-lakehouse-87813.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/how-to-architect-an-on-premises-data-lakehouse>)

Published: 2026-10-15T08:00:34Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [data lake](<https://devfeed.tech/topics/data-lake.md>), [database-optimization](<https://devfeed.tech/topics/database-optimization.md>), [acid](<https://devfeed.tech/topics/acid.md>), [Transactions](<https://devfeed.tech/topics/transactions.md>), [Cache](<https://devfeed.tech/topics/cache.md>), [control-plane](<https://devfeed.tech/topics/control-plane.md>), [parquet](<https://devfeed.tech/topics/parquet.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [air-gapped](<https://devfeed.tech/tags/air-gapped.md>), [amazon-s3](<https://devfeed.tech/tags/amazon-s3.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apache-iceberg](<https://devfeed.tech/tags/apache-iceberg.md>), [apache-iceberg-tables](<https://devfeed.tech/tags/apache-iceberg-tables.md>), [architect](<https://devfeed.tech/tags/architect.md>), [audit](<https://devfeed.tech/tags/audit.md>), [availability](<https://devfeed.tech/tags/availability.md>), [bandwidth](<https://devfeed.tech/tags/bandwidth.md>), [batch](<https://devfeed.tech/tags/batch.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [bi](<https://devfeed.tech/tags/bi.md>), [big-bang](<https://devfeed.tech/tags/big-bang.md>), [build](<https://devfeed.tech/tags/build.md>), [built-in](<https://devfeed.tech/tags/built-in.md>), [data-lakehouse](<https://devfeed.tech/tags/data-lakehouse.md>), [lakehouse](<https://devfeed.tech/tags/lakehouse.md>)

### AI overview

An on-premises lakehouse can serve sovereignty, regulatory, latency, or egress-cost needs using S3-compatible object storage, Apache Iceberg, and Trino, while separating storage from compute. Plan capacity, durability, network bandwidth, and the Iceberg catalog early; compatibility varies by provider: open-source Trino is tested with Amazon S3 and MinIO, while Starburst Enterprise documents its supported stores. The article recommends starting federation-first, then materializing frequently used paths into Iceberg, and building access controls, audit, and governed data products in from day one. Migrate from Hadoop in stages: replace Hive or Impala with Trino while HDFS remains, write new data to Iceberg, selectively convert Hive tables, then retire legacy components. Starburst Enterprise multi-cluster availability features require Elite and were in public preview in 480-e; running queries may fail during coordinator failover, so the article advises testing recovery plans that do not depend on those features.

### Source excerpt

An on-premises data lakehouse brings lakehouse capabilities into data centers you own or lease. It ...

## What Is Agentic Analytics?

DevFeed: [What Is Agentic Analytics?](<https://devfeed.tech/articles/what-is-agentic-analytics-86801.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/what-is-agentic-analytics>)

Published: 2026-10-14T08:00:33Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Agentic AI](<https://devfeed.tech/topics/what-is-agentic-ai.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [analytics stack](<https://devfeed.tech/topics/analytics-stack.md>), [metric-standardization](<https://devfeed.tech/topics/metric-standardization.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [actions](<https://devfeed.tech/tags/actions.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-control-plane](<https://devfeed.tech/tags/agentic-control-plane.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-model](<https://devfeed.tech/tags/ai-model.md>), [ai-workloads](<https://devfeed.tech/tags/ai-workloads.md>), [all](<https://devfeed.tech/tags/all.md>), [analysis](<https://devfeed.tech/tags/analysis.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [business](<https://devfeed.tech/tags/business.md>), [plan](<https://devfeed.tech/tags/plan.md>), [queries](<https://devfeed.tech/tags/queries.md>), [question](<https://devfeed.tech/tags/question.md>), [run](<https://devfeed.tech/tags/run.md>), [uses](<https://devfeed.tech/tags/uses.md>), [what-is](<https://devfeed.tech/tags/what-is.md>)

### AI overview

Agentic analytics lets an AI agent interpret a business question, plan and run queries across sources, validate results, and iterate, while a human guides the work. Unlike traditional BI, it can also take actions, which require a separately governed write path. Because agents lack analysts' implicit context, missing shared definitions can lead to plausible but wrong answers; a context layer should provide definitions, ownership, and freshness through certified data products. The post recommends starting with one domain, enforcing read-only access and audit logging at the data access layer, and evaluating results against known-correct questions. Federation can let agents query data in place, though source load and join costs may favor landing frequently queried or performance-critical data in Iceberg. MCP provides an interface, but its read-only signal is advisory; enforcement must occur server-side.

### Source excerpt

Agentic analytics uses AI agents to interpret a business question, plan the analysis, run queries ...

## What Is an Enterprise Data Strategy?

DevFeed: [What Is an Enterprise Data Strategy?](<https://devfeed.tech/articles/what-is-an-enterprise-data-strategy-86013.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/what-is-an-enterprise-data-strategy>)

Published: 2026-10-13T08:00:55Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [data-governance](<https://devfeed.tech/topics/data-governance.md>), [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [data](<https://devfeed.tech/topics/data.md>), [Data Quality](<https://devfeed.tech/topics/data-quality.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [etl](<https://devfeed.tech/topics/etl.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [data](<https://devfeed.tech/tags/data.md>), [data-strategy](<https://devfeed.tech/tags/data-strategy.md>), [enterprise](<https://devfeed.tech/tags/enterprise.md>), [enterprise-data](<https://devfeed.tech/tags/enterprise-data.md>)

### AI overview

An enterprise data strategy sets how an organization makes data accessible, governed, trusted, and accountable across systems. The article recommends starting with a few high-value business or AI use cases, then choosing centralization or federation according to access and performance needs. Federation can query data where it lives without copying it, while centralization may suit some workloads; neither alone resolves inconsistent definitions or unclear ownership. Effective strategies pair access controls and quality rules with named C-suite and business owners, shared KPIs, and a semantic layer that gives people and AI consistent meanings for business terms. The article warns that connecting AI directly to raw data or centralizing without integration work can leave context and governance problems unresolved.

### Source excerpt

An enterprise data strategy is the cross-functional plan that determines how a large organization collects, ...

## The Plane, Not the Fleet

DevFeed: [The Plane, Not the Fleet](<https://devfeed.tech/articles/the-plane-not-the-fleet-84817.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/starburst-nebula-ai-agent-the-plane-not-the-fleet>)

Published: 2026-10-08T16:37:43Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [control-plane](<https://devfeed.tech/topics/control-plane.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>), [observability](<https://devfeed.tech/topics/observability.md>), [Deployment](<https://devfeed.tech/topics/deployment.md>), [container images](<https://devfeed.tech/topics/container-images.md>), [virtual machines](<https://devfeed.tech/topics/virtual-machines.md>), [scheduling](<https://devfeed.tech/topics/scheduling.md>), [stateless](<https://devfeed.tech/topics/stateless.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentcore](<https://devfeed.tech/tags/agentcore.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-infra](<https://devfeed.tech/tags/ai-infra.md>), [atlassian](<https://devfeed.tech/tags/atlassian.md>), [audit](<https://devfeed.tech/tags/audit.md>), [audit-trail](<https://devfeed.tech/tags/audit-trail.md>), [authoring](<https://devfeed.tech/tags/authoring.md>), [bedrock](<https://devfeed.tech/tags/bedrock.md>), [bedrock-agentcore](<https://devfeed.tech/tags/bedrock-agentcore.md>), [build](<https://devfeed.tech/tags/build.md>), [building](<https://devfeed.tech/tags/building.md>), [builds](<https://devfeed.tech/tags/builds.md>), [can](<https://devfeed.tech/tags/can.md>), [ci](<https://devfeed.tech/tags/ci.md>), [engineer](<https://devfeed.tech/tags/engineer.md>), [failed](<https://devfeed.tech/tags/failed.md>), [fleet](<https://devfeed.tech/tags/fleet.md>), [on-call](<https://devfeed.tech/tags/on-call.md>)

### AI overview

Nebula's central "agent plane" handles shared infrastructure while Starburst teams own agent behavior, giving 82 agents a common deployment, identity, access-control and audit path. Agents declare reviewed tool grants; credentials stay in managed gateways, user identity is bound by the platform, and outside content is treated as data. The platform provides runtime selection, Slack routing, per-agent state and storage, scoped retrieval, scheduling and telemetry. The account also describes unresolved gaps: downstream services still see Nebula's identity, model-call paths are not yet centralized, cost figures are estimates not checked against invoices, and portfolio-wide health, deployment-blocking evaluations and drift detection remain unsolved.

### Source excerpt

Imagine you are an engineer on call, looking at a failed CI run at 2am ...

## Writing to Unity Catalog-Managed Tables with Starburst

DevFeed: [Writing to Unity Catalog-Managed Tables with Starburst](<https://devfeed.tech/articles/writing-to-unity-catalog-managed-tables-with-starburst-84821.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/writing-to-unity-catalog-managed-tables-with-starburst>)

Published: 2026-10-08T08:00:33Z

Content type: tutorial

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [Databricks Unity Catalog integration](<https://devfeed.tech/topics/databricks-unity-catalog-integration.md>), [Apache Iceberg](<https://devfeed.tech/topics/apache-iceberg.md>), [self-hosted databases vs managed services](<https://devfeed.tech/topics/self-hosted-databases-vs-managed-services.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [apis](<https://devfeed.tech/tags/apis.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [aws](<https://devfeed.tech/tags/aws.md>), [azure](<https://devfeed.tech/tags/azure.md>), [data](<https://devfeed.tech/tags/data.md>), [databricks](<https://devfeed.tech/tags/databricks.md>), [delta-lake](<https://devfeed.tech/tags/delta-lake.md>), [starburst](<https://devfeed.tech/tags/starburst.md>), [transaction](<https://devfeed.tech/tags/transaction.md>), [unity-catalog](<https://devfeed.tech/tags/unity-catalog.md>)

### AI overview

Starburst now supports reading and writing Unity Catalog-managed Delta tables, using Unity Catalog for governance and transaction commits while Starburst provides federated SQL. The article explains the architecture, configuration, supported operations, and when to choose catalog-managed or external tables.

### Source excerpt

Databricks' Unity Catalog-managed tables allow the catalog to fully own Delta Lake transaction state using ...

## Data Products: Context Engineering for AI

DevFeed: [Data Products: Context Engineering for AI](<https://devfeed.tech/articles/data-products-context-engineering-for-ai-84814.md>)

Original publisher: [Read original article](<https://www.starburst.io/blog/data-products-context-engineering-for-ai>)

Published: 2026-10-07T08:00:07Z

Content type: article

Language: en

Sources: [Starburst Data](<https://devfeed.tech/sources/starburst-data.md>)

Topics: [AI-Ready Data](<https://devfeed.tech/topics/ai-ready-data.md>), [data-engineering](<https://devfeed.tech/topics/data-engineering.md>), [semantic-layer](<https://devfeed.tech/topics/semantic-layer.md>), [context](<https://devfeed.tech/topics/context.md>), [Data science and engineering](<https://devfeed.tech/topics/data-science-and-engineering.md>)

Tags: [access-control](<https://devfeed.tech/tags/access-control.md>), [access-point](<https://devfeed.tech/tags/access-point.md>), [accessible](<https://devfeed.tech/tags/accessible.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-automation](<https://devfeed.tech/tags/ai-automation.md>), [analytics](<https://devfeed.tech/tags/analytics.md>), [apis](<https://devfeed.tech/tags/apis.md>), [app](<https://devfeed.tech/tags/app.md>), [application-development](<https://devfeed.tech/tags/application-development.md>), [business-logic](<https://devfeed.tech/tags/business-logic.md>), [centralization](<https://devfeed.tech/tags/centralization.md>), [challenge](<https://devfeed.tech/tags/challenge.md>), [context](<https://devfeed.tech/tags/context.md>), [context-engineering](<https://devfeed.tech/tags/context-engineering.md>), [data](<https://devfeed.tech/tags/data.md>), [data-products](<https://devfeed.tech/tags/data-products.md>)

### AI overview

The article argues that data products can serve as reusable context units for AI agents and as a practical way to do context engineering. It defines them as curated wrappers around datasets that bundle metadata, business logic, documentation, versioning, and access rules, making data more discoverable and usable than raw data alone. Data federation lets products combine sources where they reside instead of requiring prior centralization, and interoperable products can be composed. Interfaces such as semantic layers, data contracts, APIs, and event streams support consumption by agents, applications, and conversational interfaces. For governance, products bundle ownership and access policies: the article describes RBAC for product or catalog permissions and ABAC for finer controls such as masking PII, alongside centralized logging and lineage. It presents these as ways to support security, compliance, and auditing across distributed deployments.

### Source excerpt

AI agents are experiencing a context crisis. AI context is the talk of the industry ...

[Next page](<https://devfeed.tech/sources/starburst-data.md?cursor=WyIyMDI2LTEwLTA3VDA4OjAwOjA3KzAwOjAwIiwgIjJlMmRiOGQ1LTU2YTAtNDhmMi1iZjI5LTc1ZjhkYTI0ZDc0MiJd>)