live wire
▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation▸ThoughtSpot makes Spotter 3 default for new instances; it requires download rights and always shares query data with an LLMThoughtSpot documentation▸MotherDuck adds image classification and object location to prompt_jev(), returning typed SQL outputs with confidence scoresMotherDuck▸BigQuery conversational analytics adds Montréal and Toronto; mixed-region agents still default to globalGoogle Cloud Documentation▸Fabric enables Azure OpenAI Copilot access by default, but OpenAI-subprocessor and cross-region processing/storage switches remain offMicrosoft Learn▸Snowflake Cortex Agents now continue by default when named tools are inaccessible, emitting warnings instead of failing the runSnowflake documentation▸Databricks Genie One can use group memberships to tailor answers, while Unity Catalog remains the access-control boundaryDatabricks documentation▸Fabric data agents can query Power BI models with Read alone; Build and workspace access are unnecessaryMicrosoft Learn▸Fabric data-agent Git configs expose column flags, but selecting a table still includes every columnMicrosoft Learn▸Google Data Agent Kit offers remote Spark kernels, but notebook exports can still move results into local Pandas or ArrowGoogle Cloud Documentation▸Cube says production analytics agents need explicit partial-result states, caller-filtered model search and durable artifactsCube▸Thinking Machines names Bridgewater behind ReViSQL-K2.6, citing 92.97% SC-16 accuracy at 15% of frontier-model costThinking Machines Lab▸AWS says a mismatched data/tool/prompt release cut its Field Sales agent pass rate from 78% to 38% in pre-productionAWS Artificial Intelligence Blog▸NL2SQL Multi-Agent ships evaluation artifacts across BIRD, CSpider, Archer and KDD DataSpace—but no software licenseGitHub▸Snowflake says Cortex Analyst apps keep working, but recommends Cortex Agents; semantic views and verified queries carry overSnowflake Documentation
nl2sql.ai
analysisCloud placement

MotherDuck is coming to Azure—but today’s Copilot path still crosses into AWS

The announcement removes a future cloud-placement objection, not a present one. Teams can connect Copilot now, but should separate client integration from compute and data residency.

Current Azure queries detour to AWS; future MotherDuck runs in Azure.
AI-generated illustration
By The News Desk· Oct 17, 2026the quick take — two AI hosts go live when you do

MotherDuck says it will bring its DuckDB-based cloud warehouse to Microsoft Azure, with the first regions chosen from customer waitlist demand. It has not named a launch region, date, price or early-access scope; the company says those details will be sent before each region opens. Until then, Azure-hosted data queried through MotherDuck still reaches MotherDuck’s current AWS deployment. MotherDuck announcement

That distinction matters for teams evaluating the product through Microsoft Copilot. MotherDuck’s managed MCP endpoint already connects to Copilot Studio and GitHub Copilot, so the agent integration is available now. But an Azure client does not make the database service Azure-resident: MotherDuck’s architecture documentation currently lists six AWS regions, and an organization can exist in only one cloud region. MCP setup documentation Architecture documentation

Two decisions, not one

The practical evaluation should split into two tracks. First, choose the agent identity model. Copilot Studio can connect through OAuth, where each user authorizes a MotherDuck account, or through a shared service-account token. MotherDuck recommends a read-scaling token for a widely published agent when read-only behavior is required. Those choices determine whose permissions and audit context reach the database. MCP setup documentation

Second, decide whether today’s cloud path is acceptable. MotherDuck says its AWS service can query Parquet and CSV files in Azure Blob Storage now, but Azure charges egress when those queries cross clouds. The future Azure regions are intended to put compute beside Azure data; they are not available merely because the MCP connector works. MotherDuck announcement

Teams should therefore ask for four concrete items before treating the roadmap as a migration plan: the selected Azure region, the residency of managed storage and query compute, the migration method for an existing single-region organization, and whether identity, logs or AI-processing endpoints remain outside that region. The last question is not theoretical: MotherDuck’s current architecture notes that AI functions for its Asia-Pacific AWS regions are processed through a US endpoint. Architecture documentation

What can be tested now

A pilot does not need to wait. Teams can connect Copilot Studio with per-user OAuth, enforce a read-only token where a shared identity is unavoidable, and measure the current cross-cloud query and egress profile against Azure Blob data. That produces a useful baseline for comparing the eventual Azure deployment—but it should be labeled an AWS-hosted pilot, not evidence that Azure residency is already solved. MCP setup documentation MotherDuck announcement

sources

Filed by The News Desk. Corrections: desk@nl2sql.ai · Our standards →

comments · 0

    Comments are moderated before they appear. Your email is used once to confirm it is you — never shown, never sold. Corrections and questions get an answer from the desk when we have one.

    more from the desk