diff --git a/MobileGL/MG_State/GLState/ProgramState/ProgramLinkTask.h b/MobileGL/MG_State/GLState/ProgramState/ProgramLinkTask.h
index f18099d1..b0d0a33f 100644
--- a/MobileGL/MG_State/GLState/ProgramState/ProgramLinkTask.h
+++ b/MobileGL/MG_State/GLState/ProgramState/ProgramLinkTask.h
@@ -90,6 +90,14 @@ namespace MobileGL::MG_State::GLState {
             // the TShader's own intermediate rather than a copy. These used to die when
             // RunBody() returned, which was safe only because nothing called getIntermediate()
             // afterwards. GlslangToSpv does exactly that, so phase B has to own them.
+            //
+            // MEMORY NOTE: this is the one thing the split makes live LONGER than it used to -
+            // a glslang arena per stage, megabytes for a shaderpack, now alive from the end of
+            // phase A until phase B has generated its SPIR-V instead of dying with the link
+            // body. Phase B clears this vector as soon as GlslangToSpv returns, but a deep
+            // phase-B backlog still holds one arena per queued program. If peak RSS ever
+            // becomes the binding constraint on a pack load, THIS is the field to attack (by
+            // bounding the backlog, or by moving GlslangToSpv back into phase A).
             Vector<SharedPtr<glslang::TShader>> shaders;
             // GL enum per entry of `in.shaders`, in the same order (GetSpirvBinaryFromProgram
             // walks it to pick the intermediates).
diff --git a/MobileGL/MG_State/GLState/ProgramState/ProgramSpirvTask.cpp b/MobileGL/MG_State/GLState/ProgramState/ProgramSpirvTask.cpp
index 982d6432..bfc3943e 100644
--- a/MobileGL/MG_State/GLState/ProgramState/ProgramSpirvTask.cpp
+++ b/MobileGL/MG_State/GLState/ProgramState/ProgramSpirvTask.cpp
@@ -90,7 +90,11 @@ namespace MobileGL::MG_State::GLState {
         } const phaseAReleaser{m_phaseA};
 
         if (!m_phaseA) return;
-        const ProgramLinkTask::SpirvHandoff& handoff = m_phaseA->spirvHandoff;
+        // Non-const: the TShaders are dropped below, the moment GlslangToSpv is finished with
+        // them. This is safe by ownership rather than by locking - phase A is terminal and
+        // therefore immutable to everyone else, the GL-thread join touches only `artifacts`
+        // and `diagnostics`, and this node is the sole reader of the handoff.
+        ProgramLinkTask::SpirvHandoff& handoff = m_phaseA->spirvHandoff;
         const Uint externalIndex = m_phaseA->in.externalIndex;
         if (!handoff.ready || !handoff.reflection.program) {
             // Phase A did not reach its tail (it failed the link, or was cancelled mid-body).
@@ -105,6 +109,12 @@ namespace MobileGL::MG_State::GLState {
 
         MGLOG_D("ProgramObject %u: Starting SPIR-V generation", externalIndex);
         GenerateSpirv(handoff, externalIndex);
+        // GlslangToSpv was the only consumer of the parsed ASTs, and they are by far the
+        // largest thing this node keeps alive (one glslang arena per stage, megabytes for a
+        // shaderpack). Everything after this point works on the SPIR-V and on the TProgram's
+        // own self-contained reflection pool, so drop them here rather than at the end of the
+        // body - spirv-opt plus routing is ~87% of this node's runtime.
+        handoff.shaders.clear();
 
         MGLOG_D("ProgramObject %u: Building global-UBO routing tables", externalIndex);
         {