--- title: "Postgres 19 Release Date: Why AI Bug Hunters Pushed It Back" url: https://syntaxradar.com/radar/postgres-19-release-date-ai-delay author: "Abdul Rehman, Lead Systems Architect and Technical Director" publishedAt: 2026-11-03T00:00:00.000Z dateModified: 2026-11-03T00:00:00.000Z category: "#SYSTEMS-ARCH" topic: "postgres 19 release date" primarySource: http://rhaas.blogspot.com/2026/10/postgres-19-is-late-and-its-ais-fault.html summary: "The Postgres 19 release date slipped to November. See why AI bug reports caused it, which features were pulled, and what it means for your upgrade plan." --- # Postgres 19 Release Date: Why AI Bug Hunters Pushed It Back **Short answer:** PostgreSQL committer Robert Haas says Postgres 19 will most likely ship in November, not the usual mid-September to mid-October window, because people using LLMs are reporting far more bugs. Three major features were also pulled late in beta. **Takeaway:** Postgres 19 is a few weeks late because AI-assisted bug hunting is surfacing old defects faster than humans can review the fixes, so plan upgrades around the final release and its actual feature list. **Key figures (as reported by the source):** - Back-patched commits, Apr 8 to Sep 25 (2025 vs 2026): 724 to 1,169 - Major features reverted late in beta: 3 - Reported final release date: November 19 (reported) The Postgres 19 release date has slipped, and one of the project's own committers says AI is the reason. In a [post dated October 30](http://rhaas.blogspot.com/2026/10/postgres-19-is-late-and-its-ais-fault.html), Robert Haas, a Sr. Staff Software Engineer at Databricks and a PostgreSQL committer (one of the people allowed to merge code into the project), wrote that PostgreSQL 19 will most likely not ship until November, and that a further delay is possible. For roughly a decade, Haas writes, major releases landed no earlier than mid-September and no later than mid-October. This year the schedule broke, and three headline Postgres 19 features were pulled out late in testing. If you were planning an upgrade, or waiting for the Postgres 19 release notes to decide, the ground has shifted. The cause is a flood of bug reports from people using large language models (LLMs) to hunt for flaws, and a limited number of humans who can review the fixes. ### When is the Postgres 19 release date? Postgres 19 is coming out in November, most likely, according to Haas. He does not give a day, and he says the project could still decide to push it back further. Secondary coverage is more specific. Tech outlet [36Kr reports](https://eu.36kr.com/en/p/4035230444999817) that the final release moved from October 29 to November 19, with an extra release candidate (a near-final build published for last-round testing), RC2, added on November 12. Those dates have not been confirmed in the primary sources this article relies on, so treat the reported Postgres version 19 release date as provisional until the project announces it. How late is that? If the reported date holds, Postgres 19 lands well after the old mid-October cutoff. That is a real slip for a project known for predictable timing, but Haas calls it not a long delay in the grand scheme of things. ### What is PostgreSQL used for? PostgreSQL is used for storing and querying the structured data that applications run on. It is an open-source relational database: data lives in tables, and you ask for it with SQL. When a web app saves your account, your orders or your messages, a database like Postgres is often the thing holding them. The codebase behind it is old, large and widely relied on. Haas notes that the project's git history, converted from an older version-control system called CVS, reaches back more than 30 years, and that the first code was written a decade before that. The delay drew write-ups from database companies including [Snowflake](https://www.snowflake.com/en/blog/engineering/postgresql-19-release-delay-feature-reverts/) and Neon, where Haas's post was first published. That age is part of this story. A big, decades-old codebase holds a lot of bugs that nobody has found yet. Haas argues that is true of all software and has not stopped PostgreSQL from being popular. What changed is how cheaply those bugs can now be found. ![Postgres 19 release date timeline: April 8 feature freeze, three features reverted, the missed mid-September to mid-October window, and the reported November 12 RC2 and November 19 final release](/uploads/blog/postgres-19-release-date-ai-delay/img1.webp) ### Why AI bug reports slowed the Postgres 19 release The clearest evidence is in the commit counts Haas published. He compared the same window in two years: April 8 (feature freeze, the date after which no new features are accepted) to September 25.
April 8 to September 2520252026
Commits to the master branch1,0371,443
Commits back-patched to the new release branch724 (REL_18_STABLE)1,169 (REL_19_STABLE)
Commits also back-patched to the previous release248 (REL_17_STABLE)433 (REL_18_STABLE)
The master branch is where development happens. A back-patch is a fix copied from master into a release branch, the stable line that becomes a shipped version. Haas says commits that went into master only (not copied to a release branch) declined slightly year over year. Meanwhile, back-patched commits jumped by more than 60%, and fixes that also had to reach the previous release rose by more than 74%. Put simply, developers wrote about the same amount of new work, but spent far more of their time fixing bugs, including bugs in versions people already run in production. Haas's explanation is that a huge number of people on the internet are now reporting PostgreSQL bugs they found with LLMs. He says committers were about as conservative in 2026 as in earlier years, so the jump reflects more existing mistakes being found, not more new mistakes being made. [Phoronix's report](https://www.phoronix.com/news/PostgreSQL-19-Late) describes the same picture: a surge of AI and LLM bug reports on top of security and bug back-ports to older versions. Why does that delay a release rather than just making it better? Haas gives two reasons. LLMs can draft fixes, but those fixes are not always correct or up to PostgreSQL's standards, so humans still have to review and commit them. And time that would have gone into reviewing what was headed into version 19 got diverted to security issues and bugs in older branches, which made it slower to decide what had to be cut. ### Postgres 19 graph queries and the other reverted features Postgres 19 graph queries were one of the casualties. Between August 27 and September 15, three major features were reverted, meaning their code was taken back out of the release: - **UPDATE/DELETE FOR PORTION OF**, a way to change or delete rows for only part of a time range. - **SQL/PGQ (SQL Property Graph Queries)**, the SQL-standard way to query data as a graph of connected items. - **ALTER TABLE ... MERGE PARTITION / SPLIT PARTITION**, commands for combining or dividing the partitions (separate chunks) of a large table. Haas notes that reverting big features right after feature freeze is not especially unusual. What is unusual is pulling several of them this late in the beta period. His reading is that the same bug-fixing load left less review time, so problems with these features surfaced later than usual. If you were counting on any of these three Postgres 19 features, do not assume they will ship in this version. Check the final Postgres 19 release notes before you plan any work around them. The sources do not say when, or whether, each one will return. ![Cards showing why the Postgres 19 release slipped: back-patched commits rose from 724 to 1,169, previous-release fixes from 248 to 433, and three features were reverted](/uploads/blog/postgres-19-release-date-ai-delay/img2.webp) ### Is AI the only reason Postgres 19 is late? No, and the sources differ on how much weight to give it. Haas's own title blames AI, and his data supports a big jump in bug-fix work. But 36Kr reports that AI is not the sole cause: according to its coverage, senior developers were busy with security fixes from April to August, which also ate into release work. Haas himself names security issues in older branches as one of the two things that pulled reviewers away. The two readings overlap more than they conflict. AI tools find more bugs, some of those bugs are security problems, security fixes must be back-ported to supported versions, and every hour spent there is an hour not spent getting version 19 ready. There is also disagreement over how serious the delay is. Snowflake's write-up says the release is delayed by weeks and maybe months. Neon's version of Haas's post (it first appeared there) argues the delay is short and comes directly from the project's care for quality. Haas calls the longer-term picture less clear: if LLMs can find decades of bugs that humans missed, he asks, how does the project catch up and keep adding features? He says he does not know yet. Haas thinks the pressure reaches beyond Postgres: he writes that LLMs will require every PostgreSQL developer, and to some extent every software developer, to raise their game. For another old codebase under security scrutiny, see our look at a [recent Linux kernel vulnerability](/radar/linux-kernel-vulnerability-cve-2026-98432). ### What the delay means for your upgrade plans For most teams, a few weeks' slip moves the calendar and little else. Some practical points that follow from the facts above: - **Don't plan around a fixed date yet.** The reported Postgres version 19 release date is November 19, with RC2 on November 12, but those come from secondary coverage. Wait for the project's own announcement. - **Re-check the list of new Postgres 19 features.** Temporal updates (FOR PORTION OF), SQL/PGQ graph queries and the partition merge and split commands were all reverted. If a migration plan assumed them, it needs a rethink. - **Watch the minor releases of the version you run.** The back-patch numbers show that many fixes are landing in older branches too, not only in 19. - **Treat the delay as the project being careful.** Haas argues the delay is not particularly concerning and comes from the project's concern for quality. ### What to watch next The first thing to watch is whether the reported RC2 appears around November 12 and whether the final Postgres 19 release date holds at November 19 or slips again, which Haas says is possible. After that, watch the official Postgres 19 release notes for the final feature list, and whether the reverted features are proposed again for a later version. The longer question is the one Haas raises himself: whether a project that depends on human reviewers can keep pace once AI tools make finding bugs far cheaper than fixing them. ## Frequently asked questions ### When is the Postgres 19 release date? Robert Haas, a PostgreSQL committer, says Postgres 19 will most likely ship in November and could slip further. Secondary coverage from 36Kr reports a final release on November 19, with a second release candidate on November 12. Those dates are reported, not confirmed by the project, so wait for the official announcement before scheduling an upgrade. ### Why is the Postgres 19 release late? Haas says a huge number of people are reporting PostgreSQL bugs they found with LLMs, and humans must still review every fix. Back-patched commits rose by more than 60% year over year. Time also went to security fixes in older versions, which 36Kr reports kept senior developers busy from April to August. ### Which Postgres 19 features were removed? Three major features were reverted between August 27 and September 15: UPDATE/DELETE FOR PORTION OF, SQL/PGQ (SQL Property Graph Queries), and ALTER TABLE ... MERGE PARTITION and SPLIT PARTITION. Haas says late reverts of several big features are unusual. Check the final release notes before planning work that depends on any of them. ### What is PostgreSQL used for? PostgreSQL is an open-source relational database used to store and query the structured data behind applications, such as user accounts and orders, using SQL. Its codebase is old and large: Haas notes its git history reaches back more than 30 years, with the first code written a decade before that, which is why so many hidden bugs remain to find. ## Source Based on Robert Haas's post and Phoronix, Snowflake and 36Kr coverage. http://rhaas.blogspot.com/2026/10/postgres-19-is-late-and-its-ais-fault.html --- Published on SyntaxRadar (https://syntaxradar.com). Canonical URL: https://syntaxradar.com/radar/postgres-19-release-date-ai-delay. Last modified 2026-11-03T00:00:00.000Z.