The Open Semantic Interchange (OSI) is the most significant standardization effort the data industry has seen in a decade. It is a collaborative, open-source initiative to create a vendor-neutral specification for semantic model exchange across AI, BI, and analytics platforms. The goal is deceptively simple: define a metric once and have every tool in your stack consume it the same way.
If you are evaluating semantic layer tools, building AI agents on enterprise data, or trying to make sense of the vendor landscape around governed metrics, this guide is for you. Below is a comprehensive overview of every organization participating in the OSI initiative, what they bring to the table, and where they fit in the broader data architecture.
We have organized this guide by when each participant joined and what role they play. Think of it as the field guide to the companies shaping how your data stack will work for the next decade.
What is the Open Semantic Interchange?
OSI is an Apache 2.0-licensed, YAML-based specification for representing semantic layer constructs: datasets, metrics, dimensions, relationships, and contexts. It was announced at Snowflake's Coalesce conference in September 2025 and the v1.0 specification was published on GitHub on January 27, 2026.
The initiative solves a problem that anyone who has worked across multiple BI tools or data platforms will recognize instantly. Revenue means one thing in your finance dashboard and another in your marketing dashboard. "Active customer" has a different definition in every department. Without a shared semantic standard, every tool in your stack interprets business logic differently. And when AI agents start querying your data, inconsistent semantics don't just create confusion. They create hallucinations with authority.
The three-phase roadmap targets specification finalization (completed Q1 2026), native support in 50+ platforms with domain-specific extensions (Q2 through Q4 2026), and de facto industry standard status with a marketplace for shared semantic models (2027 and beyond).
The Initiative Leads
Four organizations co-lead the OSI initiative. They drove the initial specification and continue to shape its governance.
Snowflake
Snowflake is the AI Data Cloud company and the primary driver behind OSI. Christian Kleinerman, EVP of Product, championed the initiative from the start with a vision of open, interoperable semantic metadata. Josh Klahr, VP of Analytics Product Management, leads the OSI working group operations. Snowflake hosts the working group sessions, contributed the initial specification framework, and invested in the ecosystem through Snowflake Ventures (notably backing AtScale and Honeydew). Their Semantic Views product provides native semantic layer capabilities within Snowflake, and the Cortex Analyst feature consumes semantic models for natural language querying. Snowflake's bet on OSI is strategic: a thriving, multi-vendor semantic ecosystem makes the data warehouse more valuable, not less.
Salesforce
Salesforce co-leads OSI as one of the largest enterprise software companies in the world. Their interest is clear: Salesforce owns Tableau, one of the most widely used analytics platforms globally. Southard Jones, then Head of Product at Tableau, described OSI as a potential Rosetta Stone for business data. For Salesforce, a vendor-neutral semantic standard means Tableau can consume governed metrics defined anywhere in a customer's stack, making the platform stickier without requiring proprietary lock-in. The CRM data that lives in Salesforce is among the most semantically complex in any enterprise. Customer lifecycle definitions, pipeline stage meanings, and revenue attribution rules vary wildly across organizations. OSI gives Salesforce a path to make that complexity manageable at scale.
dbt Labs
dbt Labs is the company behind dbt (data build tool), the open-source transformation framework used by over 80,000 data teams worldwide. Their participation in OSI is foundational in a literal sense. At Coalesce 2025, dbt Labs open-sourced MetricFlow under the Apache 2.0 license and contributed it as an initial reference implementation for the OSI specification. MetricFlow powers the dbt Semantic Layer, which allows teams to define metrics in code and serve them consistently to any downstream consumer. The dbt Semantic Layer is one of the most mature implementations of the semantic layer concept, with integrations across Hex, Sigma, ThoughtSpot, and more. dbt Labs' merger with Fivetran, announced in early 2026, creates a combined entity with approximately $600 million in ARR. Semantic layer capabilities are central to the combined value proposition. For data teams already using dbt, OSI means their metric definitions become portable across any tool that adopts the standard.
RelationalAI
RelationalAI is a lesser-known but technically significant co-lead. Led by CEO Molham Aref, the company provides relational knowledge graph capabilities that run natively inside Snowflake. Their technology adds rule-based, predictive, and prescriptive reasoning to the semantic layer. Where most semantic layers handle metric definitions and dimensional modeling, RelationalAI extends this to complex business logic: constraint satisfaction, graph-based reasoning, and inference. Their participation ensures OSI can eventually handle the sophisticated enterprise use cases that go beyond simple aggregations. Think supply chain optimization rules, regulatory compliance logic, or financial risk models encoded as semantic constructs.
Founding Members (September 2025)
These organizations joined the initiative at its public launch during Coalesce 2025.
Alation
Alation is a data intelligence company that pioneered the enterprise data catalog category. Their platform combines data cataloging, data governance, and data quality capabilities. For OSI, Alation's relevance is in the metadata layer. A semantic model needs context: who owns a metric, when it was last validated, what upstream data sources feed it, who is allowed to see it. Alation provides exactly this kind of organizational and governance metadata. Their participation ensures OSI doesn't just define what metrics mean but also provides hooks for cataloging, lineage, and access governance around those definitions.
Atlan
Atlan positions itself as the active metadata platform for modern data teams. They have carved out a distinctive position as "the Context Layer for AI," winning Gartner Magic Quadrant Leader recognition in both Metadata Management and Data and Analytics Governance. Revenue grew 6x in two years. Atlan's focus goes beyond traditional data cataloging into active metadata: lineage, quality scores, usage patterns, and organizational knowledge that travels with the data. For OSI, Atlan brings the governance infrastructure that enterprises need to operationalize a shared semantic standard at scale. If OSI defines what metrics mean, Atlan helps organizations track who is using those metrics, how they change over time, and whether they remain trustworthy.
BlackRock
BlackRock is the world's largest asset manager with approximately $11.6 trillion in assets under management. Their participation in OSI is a powerful signal that the standard has enterprise-grade ambitions. Financial services firms deal with some of the most complex semantic challenges in any industry. A single concept like "net asset value" or "gross margin" can have dozens of valid definitions depending on the regulatory context, asset class, and reporting jurisdiction. BlackRock's involvement ensures the specification is tested against real-world financial complexity, not just the relatively simpler metrics found in SaaS analytics.
Blue Yonder
Blue Yonder is a supply chain management platform that serves major retailers and manufacturers globally. Supply chain is another domain where semantic consistency is critical. "Inventory coverage," "demand forecast accuracy," and "fill rate" each have multiple valid definitions depending on the business context. Blue Yonder's participation brings domain-specific semantic requirements from manufacturing, logistics, and retail. Their involvement signals that OSI is not just a BI standard. It has applications wherever business logic needs to be shared consistently across systems and teams.
Cube
Cube is an open-source semantic layer platform, originally built by Artyom Keydunov as a side project in 2018 to ensure a Slack chatbot gave consistent answers. That origin story is telling: Cube was solving the AI-semantic consistency problem before anyone called it that. Today, Cube serves as a universal semantic layer that sits between your data warehouse and any downstream consumer, whether that is a dashboard, an embedded analytics application, or an AI agent. Cube raised $25 million in June 2024 led by Databricks Ventures and rebranded as an "Agentic Analytics Platform." They earned GigaOm Leader status and have one of the most active open-source communities in the semantic layer space. For data teams looking for a vendor-neutral, API-first semantic layer, Cube is one of the strongest options.
Elementum AI
Elementum AI provides AI-powered analytics and semantic modeling capabilities focused on enterprise data. Their platform helps organizations build and manage semantic models that power AI-driven insights. While smaller than some of the other founding members, their participation in OSI reflects the growing ecosystem of specialized companies building at the intersection of semantic layers and AI applications.
Hex
Hex is a collaborative analytics workspace that combines SQL, Python, and no-code tools in a single notebook-style environment. What makes Hex relevant to OSI is their integration with the dbt Semantic Layer. Hex users can query governed metrics defined in dbt directly within their notebooks, ensuring that ad hoc analysis and data exploration use the same metric definitions as production dashboards. For organizations adopting OSI, Hex represents a critical consumption pattern: the analyst who needs to do exploratory work without breaking the semantic contract.
Honeydew
Honeydew is a YC W23 startup founded by David Krakov that takes a distinctive position in the semantic layer landscape. Their core argument is that the semantic layer should be native to the data warehouse, not a separate tool. Honeydew's platform provides a visual semantic modeling layer that generates optimized SQL and sits directly on top of cloud data warehouses. Snowflake Ventures has backed the company. For enterprises evaluating their OSI strategy, Honeydew represents the warehouse-native approach: define your semantics where your data lives, then interchange them via the OSI standard to any consuming tool.
Mistral AI
Mistral AI is a French AI company and one of Europe's leading large language model providers. They are the only major AI model provider in the OSI coalition. (OpenAI, Anthropic, and Google Gemini are notably absent.) Mistral's participation signals something important about the direction of AI. As LLMs are increasingly used to query enterprise data, the models themselves benefit from consuming governed semantic definitions rather than raw SQL or unstructured data descriptions. A semantic layer tells the LLM what "revenue" means before it tries to answer a question about revenue. Mistral's involvement ensures the OSI specification accounts for how AI models consume semantic metadata, not just how BI tools display it.
Omni
Omni is a business intelligence platform built by former Looker executives who understood that the semantic layer was Looker's real value, not the dashboards. Omni's modeling layer is heavily inspired by LookML, offering a code-first approach to defining business logic that is portable, testable, and version-controlled. Their MCP (Model Context Protocol) server integration allows AI tools like Claude and Cursor to query governed metrics directly. For teams evaluating post-Looker BI options, Omni is one of the strongest semantic-layer-first platforms. Their participation in OSI ensures the standard accounts for code-first, developer-friendly semantic modeling workflows.
Select Star
Select Star is an automated data discovery and lineage platform. They map how data flows through your entire stack: from source systems through transformations to dashboards and reports. For OSI, Select Star brings lineage context. A semantic model doesn't exist in isolation. It has upstream dependencies (which tables and columns feed this metric?) and downstream consumers (which dashboards, reports, and AI agents use it?). Select Star's technology can map these relationships automatically, which becomes essential when organizations need to assess the impact of changing a metric definition across an OSI-compliant ecosystem.
Sigma
Sigma is a cloud-native analytics platform that gives business users a spreadsheet-like interface on top of cloud data warehouses. CEO Mike Palmer has articulated the OSI value proposition clearly: organizations face inconsistent metric definitions across tools, which hampers both AI and BI adoption. OSI allows metrics to be defined once, governed centrally, and consumed universally. Sigma's spreadsheet paradigm means business users, not just engineers, interact with semantic definitions. Their participation ensures OSI considers the non-technical user experience, not just the developer workflow.
ThoughtSpot
ThoughtSpot is an analytics platform best known for its natural language search interface for data. They are a Gartner Magic Quadrant Leader with 133% year-over-year platform usage growth. In March 2026, ThoughtSpot launched "Spotter Semantics," positioning it as the industry's leading agentic semantic layer. CEO Ketan Karkhanis has been vocal about the concept of the "Agentic Semantic Layer," where AI agents route their queries through governed semantic logic rather than generating SQL directly. ThoughtSpot's aggressive positioning around agentic analytics makes them one of the most visible OSI participants. Their platform demonstrates what happens when a semantic layer is consumed not just by dashboards but by autonomous AI agents making decisions in real time.
Members Who Joined by December 2025
After the initial announcement, a wave of organizations joined during the first working group meetings in Q4 2025.
AWS
Amazon Web Services joined the OSI working group in Q4 2025. As the largest cloud infrastructure provider and the company behind Amazon Redshift, QuickSight, and a growing suite of AI services, AWS participation adds significant credibility. Redshift's semantic layer capabilities and QuickSight's BI consumption are both natural targets for OSI integration. AWS joining also signals that the three major cloud providers are now paying attention to semantic standardization, even if Google Cloud and Azure haven't formally joined yet.
Collibra
Collibra is a leading data intelligence platform focused on data governance, cataloging, and data quality. They are one of the largest players in the enterprise data governance market. For OSI, Collibra's participation is particularly important because semantic models need governance. Who approved this metric definition? When was it last reviewed? Does it comply with regulatory requirements? Collibra's platform provides the governance workflows and policy management infrastructure that enterprises need to operationalize OSI at scale. Their joining was announced alongside JPMC, signaling that OSI is moving from a vendor initiative to an enterprise and financial services concern.
DataHub
DataHub is an open-source metadata platform originally developed at LinkedIn. It provides data discovery, data governance, and data observability capabilities. DataHub's open-source roots align well with OSI's Apache 2.0 license. Their platform is widely adopted by engineering-forward organizations that want full control over their metadata infrastructure. For OSI, DataHub ensures the standard works well in open-source-first environments, not just within proprietary vendor ecosystems.
Domo
Domo is a cloud-based business intelligence and data visualization platform that specializes in bringing together data from hundreds of sources into a single platform. Their target audience is business users who need real-time operational dashboards without heavy engineering support. Domo's participation means OSI will need to support the use case of business-user-friendly consumption of semantic models, including scenarios where the same metric definitions are consumed by both self-service dashboards and AI-powered insights.
Firebolt
Firebolt is a cloud data warehouse engineered for high-performance analytics at scale. They focus on sub-second query performance for interactive analytics workloads. Their participation in OSI brings a performance-focused perspective: how does a semantic standard interact with query optimization? When metrics are defined in OSI format and consumed by a high-performance engine like Firebolt, the query plan generated needs to be efficient. This is a non-trivial technical consideration as the standard matures.
Informatica
Informatica is one of the oldest and most established names in enterprise data management. Their platform covers data integration, data quality, master data management, and data governance. For large enterprises with complex, multi-system data landscapes, Informatica is often the backbone of data operations. Their participation in OSI brings decades of experience with enterprise-scale data management challenges. Informatica understands the reality of organizations running hundreds of data sources, multiple data warehouses, and competing BI tools. That complexity is exactly what OSI needs to handle.
Instacart
Instacart is a grocery delivery and technology company that represents the enterprise customer voice in OSI. Unlike the vendor participants, Instacart is a consumer of semantic layer technology. Their involvement ensures the standard is shaped by real-world data team needs: managing thousands of metric definitions across multiple business units, serving consistent analytics to operations teams, and feeding AI models with governed data. Having an at-scale technology company participate keeps the specification grounded in practical requirements.
JPMC (JPMorgan Chase)
JPMorgan Chase is the largest bank in the United States and one of the largest financial institutions in the world. Their participation, alongside BlackRock, confirms that financial services are taking semantic standardization seriously. Banking involves regulatory reporting requirements where a single metric like "Tier 1 Capital Ratio" must be calculated identically across every reporting system. The consequences of semantic inconsistency are not just operational confusion but regulatory penalties. JPMC's involvement ensures OSI can handle the strict governance, auditability, and compliance requirements of regulated industries.
Preset
Preset is the managed cloud service for Apache Superset, one of the most widely used open-source BI and data visualization platforms. Preset provides an enterprise-grade version with additional security, collaboration, and governance features. Their participation brings the open-source BI perspective to OSI: how does a semantic standard work when the consuming analytics platform is itself open-source and community-driven? Preset ensures OSI accounts for the significant portion of the market that runs on open-source analytics infrastructure.
Starburst Data
Starburst is the enterprise company behind Trino (formerly PrestoSQL), the distributed SQL query engine. Starburst's value proposition is querying data wherever it lives without moving it: across data lakes, data warehouses, and operational databases. For OSI, Starburst brings the federated query perspective. In many enterprises, data doesn't live in a single warehouse. Semantic definitions need to work across distributed data sources, and query engines like Trino need to interpret those definitions efficiently regardless of where the underlying data is stored.
Strategy (formerly MicroStrategy)
Strategy, formerly known as MicroStrategy, is one of the original enterprise BI companies. Their semantic layer heritage goes back decades. MicroStrategy's object model, which defines metrics, attributes, hierarchies, and relationships, was arguably one of the first enterprise-scale semantic layers, predating the term itself. Their rebranding to Strategy coincides with a renewed focus on AI-powered analytics through their Mosaic product. For OSI, Strategy brings deep enterprise BI experience and the practical knowledge of what it takes to maintain semantic consistency across organizations with thousands of users and millions of report consumers.
Members Who Joined January 2026 (v1.0 Release)
The v1.0 specification release triggered a second wave of members, including several major players who initially held back.
Databricks
Databricks was notably absent from the founding coalition, maintaining its own Unity Catalog Metric Views as a competing approach. Their decision to join alongside the v1.0 release is significant. It suggests that Databricks evaluated the specification on its merits and recognized the value of contributing from inside rather than competing from outside. Databricks runs the Data + AI Summit, one of the largest data conferences in the world, and their Lakehouse platform is used by thousands of enterprise data teams. Their participation in OSI means the standard must work across both Snowflake and Databricks ecosystems, which together cover the majority of the cloud data warehouse market.
AtScale
AtScale is the company that has done more than any other to establish the semantic layer as a standalone enterprise category. Founded over 13 years ago by Dave Mariani, AtScale provides a universal semantic layer that sits between data platforms and any consuming application. They host the Semantic Layer Summit, the only major conference dedicated entirely to semantic layer technology. AtScale's path to OSI is instructive. They initially promoted their own Semantic Modeling Language (SML) as a more comprehensive approach. By January 2026, they joined, recognizing the value of contributing from inside the initiative. AtScale secured an estimated $75 to $100 million Series E in December 2025, led by Snowflake Ventures, validating the enterprise semantic layer as a standalone market category.
Qlik
Qlik is a major analytics and business intelligence platform known for its associative analytics engine. Their platform enables users to explore data freely, discovering insights through natural, associative searches rather than predefined queries. Qlik's participation brings one of the legacy BI leaders into the OSI fold. Their associative model is architecturally different from the columnar, metric-first approach of many newer semantic layers. Ensuring OSI works with Qlik's paradigm broadens the standard's applicability across the diverse BI landscape enterprises actually operate in.
JetBrains
JetBrains is the company behind some of the most popular development tools in the world: IntelliJ IDEA, PyCharm, DataGrip, and many others. Their participation in OSI is about the developer experience. As semantic models become code (YAML files, version-controlled, tested in CI/CD pipelines), the tools developers use to write and validate that code matter enormously. JetBrains could bring IDE-level support for OSI: syntax highlighting, validation, autocomplete, and refactoring for semantic model definitions. That kind of developer tooling is what turns a specification into a workflow.
Lightdash
Lightdash is an open-source BI platform built specifically for dbt users. Their product consumes dbt models and metrics natively, providing a self-service analytics layer on top of the dbt transformation workflow. For OSI, Lightdash represents the dbt-native consumption pattern: teams that define their semantic layer in dbt and need a BI tool that respects those definitions without requiring a separate modeling step. Their participation ensures the standard supports the increasingly popular "semantic layer in the transformation tool" architecture.
Coalesce
Coalesce is a data transformation platform that provides a visual, GUI-driven approach to building data pipelines and transformations in Snowflake. Where dbt is code-first, Coalesce offers a visual interface that appeals to data teams who prefer a graphical workflow. Their participation in OSI ensures the standard works well for teams that build their data foundation through visual tools, not just code-first workflows. This is important for enterprise adoption, where many data teams operate with a mix of coding and no-code approaches.
Credible
Credible is a newer entrant in the data observability and trust space. Their platform focuses on ensuring data quality and reliability across the data stack. For OSI, data quality is a foundational concern. A perfectly defined semantic model is useless if the underlying data is incomplete, stale, or incorrect. Credible's participation ensures the standard considers data quality signals as part of the semantic contract, not just the metric definition itself.
Members Who Joined February 2026
Collate
Collate is the commercial company behind OpenMetadata, an open-source metadata platform for data discovery, lineage, and governance. OpenMetadata has gained significant traction as an alternative to proprietary data catalogs. Collate's participation in OSI brings another open-source-first perspective and ensures the standard interoperates well with open metadata infrastructure. For organizations running OpenMetadata, OSI compatibility means semantic definitions can flow naturally between their metadata catalog and their BI and AI tools.
Members Who Joined March 2026
Denodo
Denodo is a global leader in data virtualization and logical data management. Their platform provides a governed semantic layer with live access to operational data across on-premises, multi-cloud, and sovereign environments without requiring data movement. Denodo's approach is philosophically different from warehouse-centric semantic layers: instead of defining metrics on top of a centralized data store, they define them on top of virtualized views that can span any data source in real time. For OSI, this is a significant architectural consideration. Denodo's participation ensures the standard works not just for warehouse-centric architectures but also for federated and virtualized data environments, where data stays in place and the semantic layer provides governed access on top.
Who is Missing?
The absences are as telling as the participants.
Microsoft is the most conspicuous gap. Power BI is the Gartner Magic Quadrant leader in analytics and its semantic model is deeply embedded in the Microsoft ecosystem. Their non-participation means OSI must prove its value without the largest BI platform on board.
SAP, IBM, and Oracle all have deep semantic layer heritage in their respective BI platforms (SAP Analytics Cloud, Cognos, Oracle Analytics) but have not joined OSI.
Among major AI model providers, only Mistral AI participates. OpenAI, Anthropic, and Google DeepMind are absent. This gap matters because the primary use case driving semantic layer urgency is AI agents consuming governed metrics. The AI model providers could benefit enormously from a standard that tells their models what business terms mean before they try to generate answers.
Google Cloud is also absent despite owning Looker, arguably the platform that originated the modern semantic layer concept through LookML. Looker's acquisition by Google for $2.6 billion in 2020 was a validation of semantic layer value. Their non-participation in OSI is a notable gap given that history.
What This Means for Your Data Strategy
If you are building a modern data architecture, the OSI participant list is a map of where the industry is heading.
The breadth of the coalition tells you that semantic standardization is not a niche concern. When Snowflake, Databricks, dbt Labs, Salesforce, AWS, JPMC, BlackRock, and 30+ other organizations agree that metrics need a common language, the question is no longer whether semantic layers matter. The question is how fast your organization adopts one.
The pattern across all 37+ participants follows what I call the Intelligence Allocation Stack. Every organization in this list is solving a piece of the same puzzle: how do you allocate intelligence to the right layer of your data architecture so that AI, BI, and human decision-making all work from the same trusted foundation?
The data foundation layer (your warehouse, your pipelines, your quality checks) needs to be solid. The semantic layer needs to translate business logic into something machines and humans can share. The orchestration layer needs to route governed data to the right tools at the right time. And the AI layer needs to consume all of this without hallucinating.
OSI is the standard that makes the semantic layer portable and interoperable. But the semantic layer alone is not enough. It needs the three other layers to work. For every dollar you spend on AI tools, six should go to the data architecture underneath them. The OSI participant list is the industry starting to realize that.
We will continue to update this guide as new organizations join the initiative and as the specification matures through Phase 2 adoption in 2026.
More from Unwind Data
Why Omni's MCP Server Changes How You Query Data in Claude and Cursor
A practitioner's guide to connecting Omni's MCP server with Claude and Cursor. How the semantic layer transforms AI analytics from demo-ready to production-grade.
The dbt Semantic Layer: How MetricFlow Turns Business Logic Into Infrastructure
The dbt semantic layer with MetricFlow lets data teams define business metrics as code. Those definitions flow to any BI tool, AI agent, or analytics platform through a single governed interface.
Semantic Interoperability: The Open Standard Connecting Your Entire Data Stack
Semantic interoperability enables different systems, tools, and AI agents to share business logic consistently. The Open Semantic Interchange (OSI) specification is the vendor-neutral standard making it happen across your entire data stack.
Ready to unlock your data potential?
Let's talk about how we can transform your data into actionable insights.
Get in touch