Microsoft adds new AI development, OneLake tools to Fabric
New features such as MCP support and simplified access to data stored across multiple clouds simplify developing agents and keep the tech giant's platform competitive.
Microsoft on Tuesday unveiled a set of new features in Fabric, including capabilities that enable users to easily access data stored in Oracle and Google Cloud platforms to better aid developing AI and analytics applications.
The new features, which are now in preview, were revealed during FabCon Vienna, a user conference for Microsoft Fabric users in Austria's capital city.
Microsoft Fabric, which was launched in November 2023, is an AI-powered data management and analytics platform. It aims to unite seven data workloads, including integration, management and analysis.
Given that the new capabilities better enable Microsoft Fabric users to access data and advance AI initiatives, they are valuable additions for the platform's users, according to William McKnight, president of McKnight Consulting Group.
"The new features in Microsoft Fabric demonstrate a significant update that enhances its capabilities as a comprehensive, enterprise-grade data platform," he said. "The additions position Fabric as a key component for customers adopting AI and undertaking data modernization initiatives."
Together, these updates represent the breadth and depth of resources Microsoft can apply to this space. Individually, none of these features is earth-shattering -- they didn't solve world hunger -- but they've touched on many of the key issues enterprises are facing when working with data.
David MenningerAnalyst, ISG Software Research
David Menninger, an analyst at ISG Software Research, similarly noted that the new features, while not significantly differentiated from those of competitors such as Oracle and Databricks, are beneficial.
"Together, these updates represent the breadth and depth of resources Microsoft can apply to this space," he said. "Individually, none of these features is earth-shattering -- they didn't solve world hunger -- but they've touched on many of the key issues enterprises are facing when working with data."
The core goal of Microsoft Fabric is to enable enterprises to unify their entire data estate, according to Arun Ulagaratchagan, corporate vice president of Azure Data.
Toward that end, Microsoft Fabric features OneLake, a data lake that enables customers to store not only their Microsoft data, but also data stored in platforms such as Amazon Redshift and Databricks. OneLake includes a capability called mirroring, which automatically generates a replica of data stored in another platform.
Mirroring enables users of multiple cloud data storage platforms to combine their data so it can be used together to inform AI and analytics applications. The latest Microsoft Fabric update adds mirroring for data stored in Oracle and Google BigQuery.
"Core to the Fabric vision … is unifying the world's data," Ulagaratchagan said. "We're methodically going after every major data source with either a shortcut, a connector or mirroring."
Meanwhile, eliminating extract, transform and load (ETL) pipelines is a significant trend in data management, according to Menninger. Mirroring makes data available without having to create costly, complex pipelines. Therefore, Menninger noted that adding mirroring for Oracle and BigQuery is significant.
"Zero copy and zero ETL are hot topics right now," he said. "Extending those capabilities to include Oracle and BigQuery will make large amounts of data much more easily accessible in the Microsoft environment, eliminating all the effort to consolidate data and maintain pipelines that perform that consolidation."
Like the added mirroring capabilities in OneLake, Graph in Fabric and Maps in Fabric aim to better unify data, according to Ulagaratchagan.
Graph in Fabric is a graph database built on graph technology from LinkedIn, which is owned by Microsoft. It uses neural networks to discover relationships between data points beyond the capabilities of traditional relational databases.
Graph databases enable data points to simultaneously connect to multiple others, while relational databases enable data points to connect to just one other at a time. Because graph databases can discover relationships that are not possible in relational databases, they are often used for fraud detection, recommendation engines and risk assessment.
Maps in Fabric enables users to add geospatial context to AI agents, generative AI assistants and analytics applications by combining streaming location data with mapping and modeling capabilities.
"The graph database was perhaps the last missing piece in Fabric becoming a complete data platform," Ulagaratchagan said.
McKnight similarly noted that the new databases -- in conjunction with the added mirroring capabilities -- make Microsoft Fabric a more complete platform for getting data ready to inform AI tools.
"These features enhance Fabric's ability to unify and enrich the data foundation [to] support the development of highly intelligent AI agents that understand business operations," he said.
MCP is an open framework developed by AI vendor Anthropic in November 2024. It simplifies developing agents by standardizing the way agents interact with the data sources that train them, including an enterprise's proprietary data and external sources, such as large language models.
With agents being the current vanguard in AI development, many data management vendors have added support for MCP. Microsoft already provides MCP support in platforms such as Copilot Studio and Azure AI Foundry. Now, it's doing so in Fabric as well.
"Fabric is the data platform for AI from Microsoft, so MCP support just extends the AI capabilities," Ulagaratchagan said.
However, despite Microsoft's emphasis on AI development in Fabric, the platform's capabilities are not yet as mature as those provided by competitors such as Databricks and Snowflake, according to McKnight.
"While Fabric is investing heavily in AI, it may not yet offer the same level of sophistication or integration of AI directly into analytical workflows as competitors," he said. "Databricks has its deep AI/ML capabilities built on an open lakehouse architecture, integrating AI directly into analytical workflows. Snowflake's … AI capabilities are maturing to offer an 'easy button' for basic AI use cases."
Next steps
Because the mirroring capabilities, graph and geospatial databases and MCP support are all in preview, they effectively comprise Fabric's roadmap.
Beyond what Microsoft has planned, McKnight suggested that the tech giant do more to create an AI development network for Fabric that extends beyond other Microsoft capabilities.
"Microsoft could deepen its commitment to open standards to foster a broader developer ecosystem, develop more differentiated advanced AI/ML capabilities and workloads and offer more granular control and performance tuning options for power users," he said. "It further could use a better position in ease-of-use for core data management and AI readiness.
Menninger noted that while Microsoft Fabric is making the notion of a data fabric an enterprise-grade reality -- data fabrics are architectures that unify an enterprise's data assets -- adding semantic modeling capabilities could be a way to continue advancing the platform's capabilities.
"Semantic models will be one of the next big frontiers," he said. "They provide context to AI and BI. However, they are not easy to create. ... I expect we'll see more from Microsoft and others over time."
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.
