WILMINGTON, Del., July 30, 2026 — EnterpriseDB (EDB) has unveiled compelling results for its EDB Postgres AI (EDB PG AI), demonstrating a distinct edge over rival AI data platforms in critical operational metrics. The independent benchmark testing, conducted by the esteemed McKnight Consulting Group and commissioned by EDB, evaluated EDB PG AI against a range of competitors, including specialized vector databases, lakehouses, document stores, and various managed Postgres platforms, all standardized on equivalent enterprise hardware. EDB PG AI emerged victorious across all tested parameters.
Performance Metrics That Matter
As enterprises increasingly pivot towards agentic AI, IDC forecasts that such technologies will account for over a quarter of global IT spending by 2029. Gartner anticipates that by 2028, one-third of enterprise applications will integrate agents, a significant leap from less than 1% just two years prior. This rapid evolution challenges the traditional architecture where AI operates on separate data copies. In response, organizations are gravitating towards Postgres as the cornerstone for AI-ready data, underscoring the importance of how Postgres is implemented.
“In every workload we assessed—ranging from raw vector searches to hybrid queries and comprehensive agent retrieval—EDB Postgres AI consistently outperformed the competition while utilizing fewer tokens and incurring lower costs,” remarked William McKnight, president of McKnight Consulting Group. “The standout feature was its reliability. Platforms that decouple storage from search exhibited significant performance degradation at scale, whereas the Postgres-based solution maintained its integrity, with EDB’s implementation being the most robust we evaluated.”
Millisecond Speed at Scale, So Agents Act in the Moment
EDB PG AI showcased the fastest query performance among all vendors evaluated by McKnight. While other platforms like lakehouses and document stores experienced latency spikes into multi-second response times, EDB PG AI consistently delivered results in milliseconds. For instance, in fraud detection scenarios, a response time of 50 milliseconds versus three seconds can be the critical difference between preventing a fraudulent transaction and incurring a loss.
- 50 ms median query latency at 50 million vectors, the fastest of any platform tested
- 80x faster than Databricks and 21x faster than MongoDB Atlas
- Up to 2x faster than the nearest managed Postgres alternatives, including Aurora and Crunchy Bridge
More Accurate Results, So Every Agent Acts on the Right Data
In terms of accuracy, EDB PG AI excelled in core vector search, achieving the highest precision across all scales tested. For autonomous agents, accuracy is paramount; the data retrieved directly influences the actions taken. A less accurate retrieval mechanism can lead to confidently executed incorrect actions at machine speed, without human oversight to intervene.
- 0.911 Recall@10, the highest accuracy for core vector search among all platforms
- 26% higher accuracy than MongoDB and 17% higher than Databricks
- Leading recall rates for any Postgres platform tested, at both 10M and 50M scales
- More accuracy than open source PostgreSQL itself, underscoring that how Postgres is run significantly impacts outcomes
End-to-End Agentic Retrieval, in One System
To evaluate the performance of enterprise agents executing multiple retrievals, McKnight modeled a concurrent three-arm agent loop involving filtered vector searches over patient vitals and clinical notes, alongside full-text searches over trial protocols. EDB PG AI completed this comprehensive loop in just 27 milliseconds, while competitors that separated storage from search lagged, often taking hundreds of milliseconds or longer and suffering accuracy losses when combining vectors and filters.
- 27ms median agent-loop latency (40ms p99) compared to 699ms for MongoDB Atlas and 5.1 seconds for Databricks
- Vector, filtered, and full-text retrieval in a single query path, utilizing one live dataset without the need for a separate vector store
- 99.7% faster write-to-read than Databricks (12 ms), ensuring the loop operates on the most current data
- 67% fewer wasted tokens than Databricks and 58% fewer than MongoDB, optimizing context retrieval
Better Price Performance
Cost efficiency in cloud computing is not merely about lower hourly rates; it also encompasses the total expense incurred to achieve desired results. EDB PG AI emerged as the most cost-effective platform in the benchmark, demonstrating superior price performance.
- 76x better price performance than Databricks and 34x better than MongoDB
- 23%–28% better price performance than the closest managed Postgres competitors when factoring in query speed
One Sovereign Foundation for Transactions, Analytics, and AI
“These results stem from a singular architectural decision: we integrate vectors, structured data, and analytics within the operational database, utilizing live data,” explained Max Romanenko, chief engineering officer at EDB. “By avoiding the establishment of a separate vector store, we eliminate the need for additional systems that require security and maintenance, as well as the risk of data becoming out of sync. Our architecture ensures that intelligence and data coexist seamlessly, which is foundational rather than an afterthought.”
EDB Postgres AI operates transactional, analytical, and agentic workloads on a unified open Postgres foundation, governed at the data layer. This architecture has earned EDB recognition as a leader in The Forrester Wave: Multimodel Data Platforms, Q2 2026. The benchmarks reinforce the notion that the manner in which Postgres is deployed can transform industry standards into significant competitive advantages.
To explore the complete benchmark report, visit: https://www.enterprisedb.com/agentic-vector-database-benchmark.
About the Benchmark
The study, conducted by McKnight Consulting Group and commissioned by EDB, evaluated pgvector performance across EDB Postgres AI, Amazon Aurora PostgreSQL, Databricks, Crunchy Bridge (Snowflake Postgres), MongoDB Atlas, and open source PostgreSQL. Each platform was normalized to equivalent enterprise-grade hardware and assessed across three workloads: core vector indexing, hybrid retrieval and filtered search, and end-to-end agentic retrieval. The full report is accessible at https://www.enterprisedb.com/agentic-vector-database-benchmark.
About EDB
EDB Postgres AI serves as the sovereign data and AI platform for the agentic enterprise. Built upon Postgres, the world’s leading open-source database, EDB Postgres AI consolidates transactional, analytical, and AI workloads into a single architecture. This integration eliminates data movement, ETL processes, and operational fragmentation, enabling enterprises to harness their data and AI capabilities on infrastructure they control—achieving production-ready sovereign AI in weeks rather than months. As a significant contributor to the PostgreSQL project, EDB is committed to the vitality of the global open-source community. For more information, visit www.enterprisedb.com.
Source: EDB
