2025 was a year of transformation, consolidation, and surprising turns in the world of databases. From major acquisitions to the rise of new standards, the landscape evolved rapidly, driven by the growing integration of AI, the maturation of open-source ecosystems, and the continued dominance of PostgreSQL. The year began with a clear signal: PostgreSQL is not just a database—it’s a platform. The release of PostgreSQL 18 in November brought key improvements, including an asynchronous I/O storage subsystem to reduce dependency on the OS page cache and support for skip scans, allowing more flexible index usage. While these features aren’t new in concept, their implementation in a mainstream, open-source DBMS marks a significant step forward. The real story, however, was the wave of activity around PostgreSQL-based services. Databricks acquired Neon for $1 billion, Snowflake bought CrunchyData for $250 million, and Microsoft launched HorizonDB, a new PostgreSQL DBaaS. These moves reflect a strategic shift: the cloud giants and data leaders are betting on PostgreSQL as the foundation for their next-generation data platforms. The most exciting development in 2025 was the emergence of two major distributed PostgreSQL projects: Supabase’s Multigres, led by Sugu, the former co-creator of Vitess and CTO of PlanetScale, and PlanetScale’s own Neki. These projects aim to bring true horizontal scaling to PostgreSQL, a long-standing challenge. The fact that Sugu’s return to the scene was met with such industry attention underscores the growing momentum behind sharded PostgreSQL systems. While past efforts like Citus, Greenplum, and Postgres-XC have had limited success, the current wave appears more focused, better funded, and backed by strong technical leadership. The year also saw the rise of the Model Context Protocol (MCP), a standardized interface that allows large language models to interact with databases through a common JSON-RPC layer. After its launch in late 2024, MCP gained serious traction in 2025, with every major database vendor—OpenAI, Google, Microsoft, Amazon, Snowflake, Yugabyte, Oracle, and more—releasing MCP servers. This has created a new layer of abstraction, enabling agents to query databases without custom code. While this opens up powerful new use cases, it also raises serious security concerns. As Andy points out, most MCP servers are simple proxies with minimal safety checks. The risk of unmonitored agents running destructive queries is real, and the industry is only beginning to grapple with automated guardrails. The file format space, long stagnant, saw a renaissance. Five new open-source columnar formats were released, with SpiralDB’s Vortex (donated to the Linux Foundation) making the biggest splash. The competition is pushing Parquet to modernize, but fragmentation remains a problem—94% of Parquet files in the wild still use v1 features from 2013. Projects like F3 and Anyblox are tackling interoperability with WASM-based decoders, offering a path to future-proofing. The next frontier may be GPU acceleration, with SpiralDB already making moves in that direction. Mergers and acquisitions were a major theme. The Fivetran and dbt Labs merger created a powerful ETL/ELT powerhouse, signaling a new era of integrated data tooling. Other notable deals included IBM’s purchase of DataStax, Salesforce’s $8 billion acquisition of Informatica, and Snowflake’s $250 million buy of CrunchyData. The trend of private equity acquiring database companies—Couchbase, SingleStore, and now Kuzu—suggests a maturing market where long-term growth is hard to achieve, and exit strategies are shifting. The year also had its share of drama. MongoDB’s lawsuit against FerretDB, a proxy that allows MongoDB applications to run on PostgreSQL, raised important questions about API replication, branding, and open-source boundaries. The irony of Microsoft donating its MongoDB-compatible DocumentDB to the Linux Foundation while suing FerretDB was not lost on the community. On the cultural front, Larry Ellison cemented his legacy—not just as a tech titan, but as the richest person in history, with a net worth of $393 billion. Oracle’s stock surge, fueled by AI infrastructure bets, propelled him to the top. Whether due to AI hype or strategic positioning, Ellison’s dominance underscores how deeply databases remain at the core of the tech economy. Finally, the year reminded us that innovation isn’t always about new systems—it’s also about resilience. Despite the closures of Hydra, PostgresML, and others, the ecosystem continues to evolve. PostgreSQL remains the beating heart of modern data infrastructure, and 2025 proved it’s far from stagnant. As we look ahead, one thing is clear: the database world is no longer just about storage and queries. It’s about intelligence, integration, and the future of data as a living, dynamic system.