October 29, 2025
Steeped in speed, served with drama
The Green Tea Garbage Collector
Go’s new memory maid speeds things up; devs cheer, skeptics ask which Google apps use it
TLDR: Go 1.25’s new “Green Tea” memory cleaner promises up to 40% less time spent tidying, and Google’s already using it. The comments cheer the clear visuals and plain-English explainer, while skeptics demand proof that big-money Google apps run on Go before they’ll buy the hype.
Go just poured a fresh cup: the “Green Tea” garbage collector in Go 1.25 promises faster apps by cleaning up unused memory more efficiently—many see about 10% faster, some claim up to 40%. It’s already running at Google and could become the default in 1.26, so the Go team wants real-world feedback now. Cue the internet sipping, slurping, and spilling tea over on go.dev.
Fans are smitten. One commenter swooned over the simple slide decks that show the old vs. new collector in action—“simply wonderful,” they said, and the thread agreed. Another called the write-up “excellent,” praising how it explains a famously mysterious system in plain English. Even the “human element” shout-out got love—rare for a post about memory cleanup! But not everyone’s ready to toast. A skeptical voice crashed the tea party to ask the money question: which profitable Google products actually run on Go? The implication: nice benchmarks, but show receipts.
Meanwhile, the nerdcore arrived with flexes and acronyms, turning the thread into a meme lounge (“VGF2P8AFFINEQB FTW”), while jokesters crowned it “garbage day canceled.” The vibe: hopeful hype with a side of “prove it in production,” plus enough tea puns to keep the kettle boiling
Key Points
- •Go 1.25 introduces the experimental Green Tea garbage collector, enabled with GOEXPERIMENT=greenteagc.
- •Workloads commonly see ~10% less time in the GC, with some reductions up to 40%.
- •Green Tea is already used in production at Google, indicating readiness.
- •The Go team plans to make Green Tea the default in Go 1.26, pending user feedback.
- •The article explains Go’s tracing GC and mark-sweep algorithm, including concurrency and parallelization aspects.