Gradle Technologies Is Now Develocity

Gradle Technologies is now Develocity, live at develocity.ai. The old gradle.com address redirects there automatically, and if you run builds with the company, hold a license, or have a support relationship, nothing on your side changes. What changed is the name and the focus.

As AI increases the volume and pace of software change, more pressure is moving to review, validation, governance, and CI efficiency. Develocity is focused on helping teams manage that shift with confidence. Develocity is the context engineering layer for software delivery, helping teams turn build, test, dependency, artifact, and CI data into actionable context for developers, platform teams, and AI agents.

Gradle Build Tool itself is unaffected. It stays open source, and is still downloaded more than 80 million times a month. It remains supported by the same team, with documentation, downloads, and community resources continuing at gradle.org. Existing builds, plugins, and workflows do not need to change.

Most CI optimizations don't survive contact with ephemeral environments

A new benchmark from the team behind Gradle Build Tool tested eleven common performance tips against ephemeral CI, the disposable, containerized runners now standard on GitHub Actions and similar platforms, and found that several stop working once persistent state disappears. Tuning incremental builds for custom tasks, for example, does almost nothing in an environment that never survives past a single run.

For teams on GitHub Actions, the free setup-gradle action already handles most of this automatically. Everyone else is left rebuilding the same caching logic by hand for every project and every version upgrade, which is the gap Develocity's Universal Cache is built to close.

Kent Beck tests prose as a programming language

In his latest Genie Lessons session, Kent Beck sits down with Dan Barrett, the founder of OpenProse, a framework that lets you write software in structured English and hand it to an AI agent to run. Beck was skeptical going in. When he told Barrett it sounded impossible, Barrett answered that people were shocked it works.

The two build a small service live: an hourly tide, weather, and moon phase lookup, working from plain-language requirements and "ensures" blocks that read like postconditions. Beck's takeaway points past the demo. The industry is moving toward specifying outcomes and letting an agent find its own path, rather than hand-holding it through each step, and multi-agent orchestration is a feature in service of that goal, not the goal itself.

AWS puts an AI agent between your pull request and production

AWS has expanded its DevOps Agent with two new preview features, Release Readiness Review and Autonomous Release Testing, that assess code changes and generate tests before a change ever merges. Release Readiness Review builds a knowledge graph of connected repositories to catch downstream failures and check changes against engineering standards written in plain language, with no policy-as-code framework required.

The move fits a pattern InfoQ has tracked across GitHub, Microsoft, CircleCI, and Dropbox this year. AI coding assistants already solved how fast code gets written, and the bottleneck has shifted to review, validation, and release. Engineering teams weighing how much autonomy to hand these agents will want to watch how the preview handles cross-repo dependencies in practice, since that's typically where automated review tools struggle.

Develocity's Artifact Cache now sits in front of JFrog Artifactory

Develocity Artifact Cache now works alongside JFrog Artifactory. Artifactory stays the system of record, while Artifact Cache adds a local caching layer, Develocity Edge, that sits beside build agents and serves dependencies over the same network instead of pulling them across the WAN on every run. On controlled benchmarks, one AndroidX project's build time dropped 88%, from nearly 50 minutes to 6, and an Apache Polaris build went from 3 minutes to 1.

The Edge nodes absorb the repeat traffic. Because each artifact is cached individually rather than as one opaque blob, a single new dependency doesn't invalidate everything else, and Artifactory only sees the small share of requests, under 5%, that are genuinely new. For large organizations, this can translate into significant operating savings and deferred repository infrastructure spend.

Provenance isn't lost in the process. Each cached artifact carries its origin, signature, and lineage, which flows into Build Scan and the supply chain dashboards in Develocity 360, and can be enforced at ingress through Provenance Governor.

The industry needs you! You might find your dream role among these job openings related to DPE, AI developer productivity, and engineering leadership.

NOTE: These postings are active at the time of sending but are subject to change.