The 2025 Semantic Layer Summit was more than a virtual event—it was a product showcase for the future of semantic modeling. AtScale unveiled its most significant updates yet, delivering innovations that make the semantic layer smarter, more interoperable, and GenAI-ready.
This blog recaps all the major product announcements from the keynote, highlighting how AtScale is advancing the modern analytics stack with performance, composability, and openness at its core.
One-Click Modeling: Instant Semantic Models from Raw Data
Semantic modeling has long been a bottleneck for analytics teams. With One-Click Modeling, AtScale uses AI to automatically generate a complete semantic model—including metrics, dimensions, relationships, and joins—based on the underlying data warehouse schema.
Key Benefits:
- Zero manual configuration
- Accelerated time to insight
- Reusable semantic objects across teams
- Designed for business users and analytics engineers alike
Composite Modeling: Composable Analytics at Scale
As organizations adopt data mesh and decentralized data product strategies, composability is no longer optional. With Composite Modeling, users can now combine, extend, and enrich existing semantic models—promoting reuse and consistency across business domains.
Use Cases:
- Blend sales, finance, and supply chain models into a unified view
- Create new cross-functional metrics like profitability or customer lifetime value
- Enable domain teams to build on governed models without duplicating logic
In-Memory Aggregates: Sub-Second Query Performance
Fast analytics shouldn’t require a trade-off in governance. AtScale’s new In-Memory Aggregates feature uses AI to cache the most frequently queried aggregates directly in memory—delivering sub-second response times for even the largest enterprise datasets.
Performance Highlights:
- Real-time analytics at enterprise scale
- Automatic tuning based on query patterns
- Works seamlessly with cloud data platforms like Snowflake and Databricks
Ecosystem Integrations: Power BI, Postgres, and More
Postgres Protocol Support
AtScale now supports the Postgres wire protocol, unlocking plug-and-play integration with tools like Superset, ThoughtSpot, and Sigma. Teams can query semantic models using native Postgres connections—no proprietary drivers required.
Deep Power BI Integration
Our enhanced Power BI connector now supports:
- Live queries (no imports)
- Native DAX support
- Model inheritance between AtScale and Power BI
- Client-side measures within governed data foundations
This makes AtScale the most seamless semantic layer integration for Power BI on the market.
GenAI + Semantic Layer: Trustworthy Natural Language Query
AtScale is the first platform to deliver production-ready GenAI experiences grounded in governed semantic models. With native integrations to Databricks Genie and Snowflake Cortex Analyst, users can ask natural language questions and receive contextually accurate answers.
What Makes It Different:
- LLMs query AtScale—not your warehouse—ensuring accurate joins and metric definitions
- No more hallucinations or inconsistent results
- Role-based access controls are enforced automatically
Open Source SML: The Standard for Semantic Modeling
As more vendors roll out proprietary semantic layers, AtScale is leading the way toward standardization with the open-source Semantic Modeling Language (SML).
SML Features:
- Written in YAML for human readability
- Object-oriented with inheritance support
- Multidimensional capabilities for enterprise-grade analytics
- CI/CD and Git friendly
AtScale’s SML ecosystem now includes:
- Prebuilt models for TPCDS, Salesforce, Snowplow, and more
- SML SDK for building models programmatically
- SML CLI tools for deployment, validation, and syntax translation
- Semantic translators for Power BI, DBT, Snowflake, and Databricks
New Ways to Build, Deploy, and Certify
To support developers and partners, AtScale also launched:
- Free Developer Edition: Start building semantic models locally—no license required
- AtScale Native App on Snowflake Marketplace: Deploy the semantic layer directly via Snowpark Container Services
- Analytics Tool Certification Program: Let ISVs validate their integrations with AtScale to ensure a trusted user experience
Why It Matters: Powering a Trusted AI + BI Future
The AtScale platform continues to evolve alongside enterprise data and AI strategies. The innovations announced at the 2025 Semantic Layer Summit reflect three core commitments:
1. Governed GenAI
AtScale transforms GenAI from risky to reliable by enforcing structure, definitions, and access controls—ensuring accurate, explainable, and consistent responses.
2. Composable, Decentralized Modeling
Composite Modeling and One-Click Modeling make it easy for teams to build, extend, and scale semantic models without losing alignment or duplicating logic.
3. Open Standards and Interoperability
SML creates a common language across platforms, unlocking cross-tool reuse, portability, and community innovation.
Get Started with AtScale
Whether you’re exploring GenAI, scaling your Power BI environment, or seeking standardization across your data stack, the AtScale semantic layer is ready for what’s next.
> Explore the SML open-source repository
> Watch the keynote on demand
SHARE
Case Study: Vodafone Portugal Modernizes Data Analytics