A new report by Forrester Research overviews 26 vendors shaping the modern data lakehouse market. The report examines how lakehouse architectures have matured into established foundations for analytics and AI, helping organizations lower infrastructure costs, promote interoperability, and support advanced analytics and AI-driven workloads at scale.

For technology and data leaders under pressure to deliver real-time analytics at scale and prepare for AI, this research provides a clear snapshot of the current market landscape and the capabilities that matter most when evaluating data lake vendors.

[CTA_MODULE]

Executive point of view

Data leaders are navigating an architecture shift, in many ways brought on by AI. Traditional data warehouses are increasingly viewed as expensive and rigid, necessarily coupling compute and storage and lacking support for unstructured data, while data lakes are reputedly difficult to govern and operationally fragile at scale. Data lakehouses aim to combine the best of warehouses and lakes, offering governance, interoperability, and affordable storage. 

Forrester’s analysis reinforces a critical reality: the lakehouse is no longer an emerging pattern — it is an established foundation for AI and real-time analytics. Organizations are consolidating analytics, data science, and AI workloads onto open, governed lakehouses to lower total cost of ownership, eliminate duplication, and accelerate time to insight.

The question is no longer whether to adopt a lakehouse, but how to operationalize it successfully without recreating the same technical debt and fragility of the past while future-proofing data platforms. Success depends on reliable, automated data ingestion, interoperability across systems, and continuous maintenance that keeps data fresh, trusted, and AI-ready.

Why data lakehouses vs. data warehouse

According to Forrester, enterprises are adopting data lakehouse architectures to simplify fragmented data ecosystems and reduce the cost and complexity of maintaining separate data warehouses and data lakes. Lakehouses combine the scalability of data lakes with the governance and performance of data warehouses, enabling analytics, AI, and real-time workloads on a single platform and addressing the limitations of legacy architectures.

Key drivers of lakehouse adoption include:

• Reducing total cost of ownership by eliminating redundant storage, compute, and ETL pipelines
• Supporting AI-ready data across structured, semi-structured, and unstructured sources
• Enabling real-time and operational analytics at scale
• Preserving flexibility through open table formats and multicloud architectures

Why data lakehouse adoption stalls in practice

While the promise of a lakehouse architecture is clear, many organizations struggle to operationalize it at scale. Common friction points include:

• High ingestion and maintenance overhead. DIY pipelines demand constant engineering effort to handle schema drift, performance tuning, and reliability issues.
• Fragmented ecosystems. Without open table formats and standardized ingestion, teams duplicate data across warehouses and tools, increasing cost and vendor lock-in.
• Delayed AI readiness. AI initiatives stall when raw, semi-structured, and unstructured data cannot be reliably ingested, governed, and accessed in real time.

Forrester highlights ecosystem fragmentation and immature semantic and data quality capabilities as a primary market challenge — reinforcing the need for platforms that reduce operational complexity rather than add to it.

What are the best use cases for data lakehouses?

Forrester identifies a set of core use cases that data lakehouses are particularly well-suited to support. These use cases share a common requirement: access to large volumes of trusted, governed data that can be used across multiple personas and workloads.

Core lakehouse use cases include:

• Business intelligence at scale, enabling high-performance analytics and reporting across the enterprise
• AI-augmented analytics, using AI to automate data preparation, discovery, and insight generation
• Data science and AI/ML, supporting model training and deployment directly on lakehouse data
• Data sharing and collaboration, enabling governed access across teams and partners
• Real-time integrated analytics, combining streaming and historical data for low-latency decision-making

Critically, Forrester emphasizes that these use cases depend not just on storage and compute capabilities, but on foundational operational capabilities such as ingestion reliability, schema management, data quality, and interoperability.

What use cases are best handled by Fivetran lakehouse support? 

While lakehouse platforms provide the analytical foundation, Forrester’s research makes clear that use cases fail when data ingestion is brittle, manual, or unreliable. Fivetran addresses this execution gap by delivering a fully managed ingestion layer that enables both core and extended lakehouse use cases.

• Business intelligence at scale: Fivetran continuously ingesting data from SaaS and operational systems into lakehouses with schema management, historical tracking, and predictable freshness. 
• AI and machine learning: Fivetran delivers current, complete, and historically accurate data directly into open lakehouse table formats. 
• Real-time and operational analytics: Fivetran supports near-real-time and operational analytics through automated, high-frequency ingestion and log-based change data capture. 
• Enterprise data hubs: Fivetran creates a managed bronze/silver data layer ensuring data and metadata to power downstream workloads.
• Customer 360: Fivetran standardizes data from hundreds of source systems into a single, governed data lake. 
• Data discovery and search: By establishing a consistent, well-structured bronze layer with automated ingestion and schema management, Fivetran makes data discoverable and trustworthy for downstream catalog, governance, and semantic tools.

How to evaluate data lakehouses 

Forrester emphasizes that the long-term success of a lakehouse strategy depends less on selecting a specific platform and more on evaluating operational capabilities that determine reliability, trust, and scalability.

Key evaluation criteria include:

• Open interoperability. Iceberg and Delta Lake formats ensure compatibility with any analytics or AI engine, preventing vendor lock-in.
• Governed data foundations. ACID-compliant tables, metadata management, and lineage support regulatory and operational requirements.
• Continuous ingestion at scale. Automated, no-code pipelines keep data fresh across structured, semi-structured, and unstructured sources.
• AI-ready architecture. Reliable access to raw and historical data enables faster model training, experimentation, and deployment. Forrester warns that without strong execution across these areas, organizations risk recreating the same operational debt that undermined previous data platforms—even when adopting modern lakehouse technologies.

Why Fivetran is a data lakehouse leading vendor

This report shows that data lakehouse vendors differ significantly in focus, deployment models, and strengths. Data leaders evaluating lakehouse solutions should prioritize:

Automation. Fivetran minimizes the need for engineering time wherever possible. This includes automated data extraction, loading, and transformation, as well as cleaning, deduplicating, compacting, partitioning, and clustering data.

• Fully managed, automated data extraction and loading from over 700 SaaS and operational sources
• Automated schema detection, schema evolution, and historical change capture (CDC)
• ELT-based transformations executed directly in the destination
• Built-in deduplication and data normalization
  Continuous pipeline monitoring with self-healing behavior

Open table standards. Bringing data warehouse-like functionality to data lakehouses Fivetran avoids lock-in by standardizing on open table formats. Open table standards can also be accessed by a wide range of query engines, allowing multiple data teams to use the same data architecture for different use cases as needed.

• Native ingestion into open cloud storage (Amazon S3, Azure Data Lake Storage, Google Cloud Storage)
• Automatic creation and maintenance of Apache Iceberg and Delta Lake metadata
• Compatibility with multiple query engines and compute platforms
• Decoupled storage and compute for architectural flexibility

Reliability at scale. Fivetran pipelines can support real-time and AI workloads without constant firefighting. Key features and capabilities we include are support for performance and scale optimization, data catalogs, data lineage, data governance, and all forms of uptime and data quality assurance.

• Log-based change data capture for high-volume, high-velocity sources
• High-frequency and near-real-time sync capabilities
• Automatic handling of schema changes without pipeline failures
• 99.9% reliable data delivery

Future-proof architecture. The Fivetran platform offers a unified data architecture can support evolving analytics and AI use cases without replatforming. Support for schema evolution and metadata management is essential as well.

• Unified ingestion architecture across warehouse, lake, and lakehouse destinations
• Support for structured, semi-structured, and unstructured data
• Comprehensive metadata capture to support lineage and governance
• Open, extensible architecture aligned with modern lakehouse standards

Fivetran Managed Data Lake Service provides a fully managed, automated ingestion layer that enables modern data lake and lakehouse use cases—without operational complexity.

Conclusion

The lakehouse market has reached maturity, but operational success is still not guaranteed. Forrester’s research makes it clear that organizations that simplify ingestion, embrace open architectures, and reduce maintenance burden are best positioned to realize faster insights, lower costs, and AI-driven innovation. Fivetran Managed Data Lake Service future-proofs your architecture by making the lakehouse operational, enabling teams to scale analytics and AI on trusted, continuously updated data.

[CTA_MODULE]