Publications

Research Reports

Projecting Analytic Data Workloads for Agentic AI - WinterCorp Report

This report examines how agentic AI changes the scale and nature of workloads for the analytic data platform. As enterprises implement their agentic AI strategies at scale, many will need to reconsider their data platform(s). Many will either need to install an additional data platform or replace the one they are using with a platform able to handle these new, dramatically different workloads.

Why? Because agents – especially when operating at high volume or in large numbers – will generate unprecedented numbers and types of interactions with the database. Most analytic databases – whether cloud, on prem or hybrid – were not designed for such workloads. This report describes the nature of these changes and recommends an approach to preparing for — and projecting — the workloads that most of us will be facing in the months and years ahead

Research Notes

Oracle AI Database and Agentic AI - Secure, Practical Agentic AI for the Enterprise

WinterCorp’s research note examines Oracle’s major Agentic AI announcements from the AI World Tour in London, where Oracle unveiled a suite of enterprise-class capabilities that analyst Richard Winter argues are meaningfully different from what the rest of the industry has offered. From a no-code agent-building platform to novel approaches to data security and answer trustworthiness, the note explores how Oracle is tackling the real obstacles that have been keeping AI projects from reaching production — and why that represents a significant opportunity for Oracle customers.

Research Notes

“You’re Gonna Need a Bigger Boat” — A Metaphor and a Recommendation for AI and the Data Platform:

This note argues that enterprise AI will radically expand who touches data (from thousands to potentially tens of thousands plus autonomous agents), drive heavier and more complex workloads (natural language querying, semantic search, multimodal analytics), and break data silos, pressuring architectures built for the BI era rather than AI. The prescription is to run a quantitative assessment of your current data platform against anticipated AI workloads, then act on one of four outcomes: no change, optimize, add a dedicated AI engine (for example via open table formats), or fully migrate, so AI rollouts do not fail for lack of data platform readiness.

Research Notes

Oracle AI World - Major Advances in AI for Enterprise Data

WinterCorp’s research note highlights Oracle’s comprehensive AI strategy unveiled at AI World, positioning Oracle as a leader in enterprise-grade AI for data management. Centered around the new Oracle AI Database 26ai, Oracle integrates vector data, similarity search, generative AI capabilities, and advanced SQL extensions directly into its converged database architecture, enabling combined structured and vector search across diverse data types. Beyond the database, Oracle announced AI enhancements across its Autonomous AI Lakehouse, AI Data Platform, OCI, Exadata, and multi-cloud partnerships, delivering flexible deployment across cloud, hybrid, and on-prem environments. WinterCorp concludes that Oracle stands out for offering secure, scalable, and cost-optimized AI solutions designed for large-scale, mission-critical enterprise workloads, empowering organizations to deploy production-ready AI solutions grounded in enterprise data .

Research Notes

The Data Warehouse/Lakehouse as a Digital Skyscraper

In this thought-provoking note, Richard Winter likens large-scale data platforms to digital skyscrapers, arguing that enterprises operating at scale must shift from tactical vendor selection and adopt a strategic, engineering-first approach. While many companies can succeed with common platforms, those with enterprise-class needs—high performance, mission-critical workloads, large data volumes, or ambitious AI goals—must focus on foundational, quantitative factors like scalability, service level delivery, cost of operation, and security. Winter warns against making platform choices based on brand popularity or developer preference and emphasizes the need for early measurement, testing, and strategic alignment to avoid costly failures.

Research Notes

Starburst Lakeside AI – AI Empowered by All of Your Data – ACAN Research Note

In this June 2025 ACAN Research Note, Richard Winter highlights Starburst’s Lakeside AI, which enables generative, predictive, and agentic AI directly on distributed enterprise data without requiring centralization. Built on the open-source Trino engine, Starburst provides federated access to over 50 data sources—across cloud, on-prem, structured and unstructured environments—through a single query interface. New features like AI Agents, AI Workflows, governance tools, and vector search allow real-time, governed AI across fragmented data estates. Starburst’s architecture empowers organizations to rapidly deploy scalable AI applications without data movement, making it a strategic solution for complex, distributed enterprises seeking to unlock AI’s full potential.

Research Notes

AI, Data and the Enterprise

This WinterCorp research note outlines Oracle’s comprehensive AI strategy, centered on its converged Oracle Database 23ai. The platform supports GenAI and predictive AI, integrating vector search, text-to-SQL, multi-model data (relational, JSON, graph, vector, etc.), and Oracle’s globally distributed and autonomous capabilities. Innovations like JSON and Graph Duality Views, True Cache, and support for open table formats like Iceberg enhance flexibility and data integrity. Oracle’s AI infrastructure spans all major public clouds, on-prem, and hybrid deployments—boosted by NVIDIA hardware. WinterCorp recommends Oracle Database 23ai for enterprises with mission-critical data needs, citing its unified, scalable, and AI-ready approach.

Research Notes

Data That is Trusted by Users – ACAN Research Note

This ACAN Research Note by Richard Winter spotlights Validatar, a data quality automation platform that has enabled organizations like Texas Mutual Insurance to achieve unprecedented trust in their data. Validatar automates testing and monitoring across the full data lifecycle, using a data quality catalog and metadata driven, rule-based logic to detect issues before they reach end users. With support for complex architectures like DataVault 2.0 and compatibility across data repositories, Validatar allows for scalable, transparent, and effective quality management. WinterCorp recommends Validatar for organizations committed to data quality, noting the significant business value in delivering data users can rely on.

Research Notes

Oracle CloudWorld 2024

This WinterCorp Research Note, authored by Richard Winter, highlights Oracle’s major announcements at Oracle CloudWorld 2024. Key among them is the availability of Oracle Autonomous Database on all major hyperscale clouds (AWS, Azure, Google Cloud, and Oracle Cloud), marking a significant shift in cloud accessibility and performance. Additional innovations include Exascale architecture, which enables full separation and elasticity of compute and storage, and advanced GenAI capabilities like GenDev, AI Vector Search, and RAG support. Oracle also unveiled Graph Duality Views, allowing simultaneous relational and graph views of data from a single source. WinterCorp recommends customers that need versatility, and innovation leadership, take a look at Oracle’s converged platform for their enterprise-grade cloud database.

Research Notes

The HRBR Enterprise Data Exchange Platform

HARBR (HARBR Group Ltd.) is a cloud-based data commerce company that provides a platform for creating, managing, and distributing data products. This note discusses how the platform discards old ideas and enables business breakthroughs via a collaborative use of data and analytics, and advanced ideas about the value, productization and monetization of data.

Research Notes

Oracle Introduces JSON Relational Duality

This WinterCorp Research Note highlights Oracle’s innovative JSON Relational Duality feature in Oracle Database 23c, which allows data to be stored once yet accessed simultaneously as both JSON documents and normalized relational tables. This capability resolves the long-standing object-relational mismatch by enabling seamless dual access—via document APIs or SQL—while maintaining ACID compliance and avoiding data duplication. Users can create multiple JSON views of the same data, facilitating flexible application development. WinterCorp recommends that organizations dealing with document data take a look at JSON Relational Duality, noting its unique ability to simplify development and boost data integrity and efficiency across diverse use cases.

Research Notes

The InterSystems Speed Test

This WinterCorp report evaluates InterSystems IRIS against leading cloud databases using the InterSystems Speed Test, measuring insert rate, query rate, data latency, and scalability. In both single-node and 1–4 node clustered configurations on AWS, IRIS consistently outperformed AWS Aurora MySQL, Oracle, PostgreSQL, MariaDB, and SQL Server. It achieved up to 9x faster inserts and 600x faster queries, with much lower data latency and better scaling efficiency per vCPU. IRIS’s in-memory distributed cache architecture enables immediate query visibility post-insert, unlike Aurora MySQL’s disk-based replication which introduces lag. WinterCorp recommends InterSystems IRIS for enterprises needing high-throughput, low-latency transactional-analytic performance in the cloud.

Research Notes

A Game Changing Platform for Cloud Data Analytics

This WinterCorp report highlights Teradata’s VantageCloud Lake with ClearScape Analytics as a breakthrough platform in cloud data analytics. It details how the platform combines full cloud-native architecture with enterprise-grade analytics capabilities, offering shared low-cost object storage, elastic multi-cluster compute, centralized administration, and in-database AI/ML analytics. ClearScape Analytics supports scalable, real-time, and complex query processing while ensuring cost-efficiency and business autonomy. WinterCorp calls VantageCloud Lake a uniquely powerful solution for demanding modern day workloads, and recommends that customers seeking both, agility and high-performance analytics in the cloud, take a look at it.

Research Notes

Unified Analytics with Open Data – Dremio – ACAN Research Note

In this May 2024 research note, Richard Winter explores Dremio’s second-generation lakehouse platform, which uses open data formats—especially Apache Iceberg—to deliver high-performance, flexible analytics without vendor lock-in. Built on Apache Arrow, Dremio offers a massively parallel SQL engine, patented query acceleration (Reflections), semantic layers for self-service, and virtual data versioning for experimentation and governance. It supports hybrid, multi-cloud, and on-prem deployments, and is praised for ease of use, reduced cloud costs, and agility. WinterCorp recommends Dremio as a serious option for organizations seeking open-format analytics solutions and faster time to value, especially at scale.

Research Notes

“You’re Gonna Need a Bigger Boat” — A Metaphor and a Recommendation for AI and the Data Platform:

This note argues that enterprise AI will radically expand who touches data (from thousands to potentially tens of thousands plus autonomous agents), drive heavier and more complex workloads (natural language querying, semantic search, multimodal analytics), and break data silos, pressuring architectures built for the BI era rather than AI. The prescription is to run a quantitative assessment of your current data platform against anticipated AI workloads, then act on one of four outcomes: no change, optimize, add a dedicated AI engine (for example via open table formats), or fully migrate, so AI rollouts do not fail for lack of data platform readiness.

Research Notes

“You’re Gonna Need a Bigger Boat” — A Metaphor and a Recommendation for AI and the Data Platform:

This note argues that enterprise AI will radically expand who touches data (from thousands to potentially tens of thousands plus autonomous agents), drive heavier and more complex workloads (natural language querying, semantic search, multimodal analytics), and break data silos, pressuring architectures built for the BI era rather than AI. The prescription is to run a quantitative assessment of your current data platform against anticipated AI workloads, then act on one of four outcomes: no change, optimize, add a dedicated AI engine (for example via open table formats), or fully migrate, so AI rollouts do not fail for lack of data platform readiness.

Research Notes

“You’re Gonna Need a Bigger Boat” — A Metaphor and a Recommendation for AI and the Data Platform:

This note argues that enterprise AI will radically expand who touches data (from thousands to potentially tens of thousands plus autonomous agents), drive heavier and more complex workloads (natural language querying, semantic search, multimodal analytics), and break data silos, pressuring architectures built for the BI era rather than AI. The prescription is to run a quantitative assessment of your current data platform against anticipated AI workloads, then act on one of four outcomes: no change, optimize, add a dedicated AI engine (for example via open table formats), or fully migrate, so AI rollouts do not fail for lack of data platform readiness.

Research Notes

“You’re Gonna Need a Bigger Boat” — A Metaphor and a Recommendation for AI and the Data Platform:

This note argues that enterprise AI will radically expand who touches data (from thousands to potentially tens of thousands plus autonomous agents), drive heavier and more complex workloads (natural language querying, semantic search, multimodal analytics), and break data silos, pressuring architectures built for the BI era rather than AI. The prescription is to run a quantitative assessment of your current data platform against anticipated AI workloads, then act on one of four outcomes: no change, optimize, add a dedicated AI engine (for example via open table formats), or fully migrate, so AI rollouts do not fail for lack of data platform readiness.

© 2026 WinterCorp. All Rights Reserved.