You've spent years close to a broken industry or process. You know exactly what would fix it. But you’ve never had the time or resources to make it happen.

With Lovable, your expertise is all you need. Describe the solution to your problem — a website, a dashboard, an internal tool — and get professional-grade software back. You make every call on how it works, because you're the one who knows what it needs to do.

Start today and take $5 off your first month.

Build the fix, $5 off

_{^{Discount of $5 applies to first month of any paid plan. You need to sign up via this webpage in order to claim the discount. The discount applies to new and returning users without an active paid plan. Terms apply.}}

Something a little different this week.

I combed through 997 headlines just now, and quite honestly, I did not see any hard hitting news for the last 7 days that I think is noteworthy and impactful.

I have to be inspired every week to cover the news, but for some strange reason, I did not see anything of note. A lot of thought is put into the curation of the newsletter. I have stated many times that this will not be an AI newsletter masquerading as a database platform/cloud database newsletter. I am not going to add AI filler to the newsletter just because I can.

I have to see news that is impactful to my readers and myself. There is usually a vast amount of news stories and a lot of important headlines that you expect to see. I’m just not going through the motions to churn out the regularly formatted newsletter this week.

However, on Thursday October 15, I had the inspiration to jot down about 11 things that I am seeing in the database world that have stuck with me for the last few weeks.

These could be things I see and research or things that I discuss with my colleagues at work. They could also be from the from conversations with vendors and implementers, or from some of the previous news stories that I covered in the last 6 months.

Here are said 11 issues/points that I made note of this week, that I think are important. I will have my regular installment of the twice yearly State of the Union late December.

Here we go.

Postgres everywhere + getting swallowed up.

Postgres has stopped being a database you choose and become the interface everybody else builds behind, around and underneath. Snowflake bought Crunchy Data, Databricks turned Neon into Lakebase, and agents now create 70% of Supabase's new databases, so the elephant keeps winning while the companies built around it keep getting swallowed up.

The open question is not whether Postgres wins, but who ends up owning the Postgres your applications actually run on.

Agent-controlled databases.

One trend I keep seeing is the rise of databases that can be provisioned, queried, modified, and even administered by AI agents with increasingly less human involvement.

Supabase, Databricks, MongoDB, and others are moving in this direction, and I suspect we are only seeing the beginning of what will become a much larger architectural shift. What I am particularly interested in is how organizations intend to govern all of this, because giving an AI agent the ability to make database changes is one thing, but holding somebody accountable when things go wrong is quite another.

Multi-model databases with various engines.

For decades, we have largely categorized databases according to their particular strengths, whether relational, document, graph, columnar, or more recently vector databases. Now I am seeing vendors combine multiple data models and specialized execution engines within what appears to the customer as a single database platform, with AWS incorporating DuckDB capabilities into Aurora PostgreSQL being a particularly interesting example.

To me, this is a significant architectural development because we may be moving away from choosing one database engine to perform every task and toward platforms that intelligently use different engines for different workloads.

Table formats still competing.

I keep hearing that Apache Iceberg has won the open table format war, but I am not quite ready to declare victory when Delta Lake continues to evolve and vendors are still competing over catalogs, metadata, and interoperability.

What I find interesting is that the conversation has moved beyond simply storing data in an open format to determining which platform controls the metadata, governance, and execution capabilities surrounding that data. Having spent years dealing with proprietary database technologies, I welcome open table formats, but I remain skeptical of any vendor claiming openness when moving away from its platform can still be a major undertaking.

The vector database still in flux.

A few years ago, dedicated vector databases were being positioned as an essential component of the emerging AI technology stack, and companies like Pinecone, Weaviate, Zilliz, and others attracted considerable attention.

Fast-forward to today, and vector search is increasingly available inside established relational, document, and cloud databases, which raises the obvious question of how much room remains for standalone vector database vendors. I do not think these specialized platforms are going away, particularly for demanding workloads, but I suspect we will see further consolidation as vector search becomes a standard database capability rather than something exotic.

Looking to reach CTOs, CIOs, and enterprise Data Engineers and Data Architects?

Limited sponsorship slots available each month.

👉 Sponsor Cloud Database Insider

Fabric still in third place.

Microsoft Fabric continues to push hard into the enterprise ecosystem, but in terms of core architectural adoption for heavy analytical workloads, it still trails behind Snowflake and Databricks.

While its integration with Power BI and Microsoft Office makes it attractive to business units, enterprise data teams are taking a more cautious approach when migrating core data warehouse pipelines. It remains a formidable challenger, but Snowflake and Databricks retain a firm grip on top-tier data platform mindshare.

ClickHouse as an alternative, but underrated.

ClickHouse is one of those database platforms that keeps appearing in my research, particularly in real-time analytics, observability, high-performance querying, and increasingly AI-related workloads. I am not suggesting that ClickHouse is going to replace SQL Server, PostgreSQL, Snowflake, or Databricks across the board, but I do think it deserves far more consideration when organizations are evaluating alternatives for demanding analytical workloads.

One thing I have learned over the years is that some of the most interesting technologies are not necessarily the ones with the biggest marketing budgets or the most recognizable names.

There will always be fiefdoms and people who like what they like despite alternatives.

After nearly three decades of working with databases and data platforms, one thing I have learned is that technology decisions are not always made for purely technical reasons. Organizations have their fiefdoms, established loyalties, preferred vendors, and people who will defend a particular technology long after compelling alternatives have emerged. I have seen enough of this throughout my career to believe that no amount of innovation, open standards, or impressive benchmarks will completely eliminate the human and political dimensions of enterprise technology selection.

As you get more senior, the job becomes less about picking the right database and more about diplomacy, because the architecture diagram never shows the org chart that actually decides it.

AI is not making data management any easier.

For all the excitement surrounding AI and the extraordinary number of products claiming to revolutionize data management, I am not convinced that managing enterprise data has become any easier. We still have the same fundamental problems involving data quality, inconsistent business definitions, governance, lineage, security, integration, and the enormous complexity of keeping enterprise systems running reliably. In fact, I would argue that introducing AI into an already complicated data environment can create another layer of complexity unless organizations have first addressed the underlying architectural and data management problems.

I keep returning to boring database words like idempotency and governance, because the work has not disappeared, only moved closer to production.

I personally do not see agentic AI in my day-to-day with respect to data management, reporting, and action taken on insights from data observability.

I read an extraordinary number of announcements about agentic AI autonomously managing databases, orchestrating pipelines, responding to observability alerts, and taking corrective action without human intervention, but I have to say that I am not seeing much of this in my own day-to-day work.

In the enterprise environments I deal with, we are still relying on dashboards, monitoring tools, operational processes, and people to investigate problems, interpret data, and determine what actions should be taken. I am certainly not dismissing the potential of agentic AI, but I think there is a considerable difference between what vendors demonstrate and what organizations are actually prepared to put into production.

Still waiting for the Databricks IPO.

Like a lot of you, I am still waiting for the Databricks IPO. As of mid-October there is still no S-1, public or confidential, and Ali Ghodsi said back in June that this is a bad year to go public, which points to 2027 or later. With private capital this easy to come by, I am starting to wonder whether Databricks needs the public markets at all, or whether I will be writing this same line again next October.

Gladstone Benjamin

🎯 Level Up Your Data Career: Database Certification Strategy Call

Struggling to figure out which cloud database certifications actually move the needle for your career and salary?

Skip the guesswork. Book a 1-on-1, 45-minute Database Certification Strategy Call directly with Gladstone Benjamin. Drawing from 27+ years as a Data Architect and DBA and holding multi-cloud certifications across AWS, Azure, GCP, OCI, Snowflake, and Databricks, I’ll help you map out a targeted, high-ROI certification roadmap tailored to your specific background and goals.

Special offer for Cloud Database Insider readers: Save 25% OFF your session!

👉 Book Your Certification Strategy Call Here (Use promo code SEPTCERT2026 at checkout)