diff --git a/src/v2/extdeps/languages/dag.dag b/src/v2/extdeps/languages/dag.dag
index 24d62c1b2b1..ef46fc7c9b7 100644
--- a/src/v2/extdeps/languages/dag.dag
+++ b/src/v2/extdeps/languages/dag.dag
@@ -1219,9 +1219,13 @@ fn dag_grammar_stmt_seq_expr() -> GrammarExpr {
   dag_grammar_repeat(element: dag_grammar_nonterminal(production: ^dag_production_stmt))
 }
 
-// AN EXPRESSION STATEMENT NEVER BEGINS WITH A LET SPELLING. `name =` and `node name` head the two
-// let_sugar forms (dag_grammar_let_sugar_expr) and begin no expression, so the expression
-// alternative is guarded by exactly those heads: the two alternatives are disjoint by construction
+// AN EXPRESSION STATEMENT NEVER BEGINS WITH A LET SPELLING. `name =` and `node name :` / `node name =`
+// head the two let_sugar forms (dag_grammar_let_sugar_expr) and begin no expression, so the
+// expression alternative is guarded by exactly those heads. The node head is all three tokens:
+// a bare `node` statement followed on the next line by a statement that begins with a name (the
+// next match arm, an `if`) begins `node name` too, and it is an expression; a two-token guard
+// refused it while the sugar, lacking its `:` / `=`, refused it as well. So the guard is read at
+// depth three: the two alternatives are disjoint by construction
 // and the choice-overlap proof reads the guard. The guard captures nothing; the statement's tree is
 // the expression's.
 fn dag_grammar_let_sugar_head_expr() -> GrammarExpr {
@@ -1232,7 +1236,13 @@ fn dag_grammar_let_sugar_head_expr() -> GrammarExpr {
     ),
     right: dag_grammar_sequence(
       left: dag_grammar_literal_terminal(token_class: ^dag_token_ident, lexeme: symbol_intern_lexeme(lexeme: "node")),
-      right: dag_grammar_binding_name_terminal()
+      right: dag_grammar_sequence(
+        left: dag_grammar_binding_name_terminal(),
+        right: dag_grammar_choice(
+          left: dag_grammar_terminal(token_class: ^dag_token_colon),
+          right: dag_grammar_terminal(token_class: ^dag_token_eq)
+        )
+      )
     )
   )
 }
diff --git a/src/v2/test/claim/parse/bare_node_statement_route_test.dag b/src/v2/test/claim/parse/bare_node_statement_route_test.dag
new file mode 100644
index 00000000000..59b41df596c
--- /dev/null
+++ b/src/v2/test/claim/parse/bare_node_statement_route_test.dag
@@ -0,0 +1,86 @@
+module v2.test.parse.bare_node_statement_route
+
+import v2.compiler.parse { parse_module_prepared }
+import v2.compiler.program_assembly { dag_prepared_grammar }
+import v2.compiler.tokenize { tokenize }
+import v2.extdeps.languages.dag { dag_lex, parse_production_emitted_identity_optional }
+import v2.std.diagnostic { Accepted, Outcome, Rejected, bind_outcome }
+import v2.std.live_tree { LiveTreeDisposition, SubstrateInputsOnly }
+import v2.std.logic { Bool }
+import v2.std.node { Node, Symbol }
+import v2.std.optional { Absent, Present }
+
+data live_tree_disposition: LiveTreeDisposition = SubstrateInputsOnly
+
+// A BARE `node` STATEMENT IS AN EXPRESSION, EVEN WHEN THE NEXT LINE BEGINS WITH A NAME
+// (v2.extdeps.languages.dag dag_grammar_let_sugar_head_expr). The node let-sugar head is
+// `node name :` / `node name =`; a two-token guard `node name` also excluded a bare `node` value
+// followed by the next match arm or an `if`, and five corpus files refused (among them
+// src/v2/std/node.dag). The shapes are the regression's own: an arm `=> node` before the next arm,
+// and `node` before an `if`. Each takes the expression route (no let_expr); the sugar `node y = 1`
+// still takes the let route (one let_expr). The first two sources are the regression's fixtures
+// verbatim (a1, a3). Each claim pays ONE parse of the smallest module that
+// reaches its route.
+data bns_arm_source: String = "module a1\ntype O = S { node: Int } | L { x: Int }\nfn f(o: O) -> Int {\n  match o {\n    S { node: node } => node\n    L { x: x } => x\n  }\n}\n"
+data bns_if_source: String = "module a3\nfn f(node: Int, c: Bool) -> Int {\n  node\n  if c {\n    1\n  } else {\n    0\n  }\n}\n"
+data bns_sugar_source: String = "module p\nfn f() -> I {\n  node y = 1\n  y\n}\n"
+
+fn bns_parse(text: String, file: Symbol) -> Outcome<Node> {
+  match dag_prepared_grammar() {
+    Rejected { diagnostics: d } => Rejected { diagnostics: d }
+    Accepted { value: prepared, diagnostics: residue } =>
+      bind_outcome(
+        o: tokenize(text: text, file: file, rules: dag_lex()),
+        f: fn(token_stream) {
+          match parse_module_prepared(tokens: token_stream, prepared: prepared, validation_residue: residue) {
+            Rejected { diagnostics: r } => Rejected { diagnostics: r }
+            Accepted { value: artifact, diagnostics: d } => Accepted { value: artifact.tree, diagnostics: d }
+          }
+        }
+      )
+  }
+}
+
+fn bns_lets(root: Node) -> Int {
+  let here = match parse_production_emitted_identity_optional(node: root) {
+    Present { value: id } => if id == ^dag_surface_let_expr { 1 } else { 0 }
+    Absent => 0
+  }
+  fold(root.children, init: here, f: fn(acc, e) { acc + bns_lets(root: e.target) })
+}
+
+// Accepted, holding exactly `lets` let_expr nodes.
+fn bns_route(text: String, file: Symbol, lets: Int) -> Bool {
+  match bns_parse(text: text, file: file) {
+    Accepted { value: tree, diagnostics: _ } => bns_lets(root: tree) == lets
+    Rejected { diagnostics: _ } => false
+  }
+}
+
+// THE THREE ROUTES, FROM ONE PARSE OF EACH SOURCE. Nullary and pure, enrolled WARM in
+// v2.workflow.floor_pure_producer_share: each claim reads its stored Bool.
+type BnsVerdicts {
+  arm_before_arm: Bool
+  node_before_if: Bool
+  sugar: Bool
+}
+
+fn bns_verdicts() -> BnsVerdicts {
+  BnsVerdicts {
+    arm_before_arm: bns_route(text: bns_arm_source, file: ^bns_arm, lets: 0),
+    node_before_if: bns_route(text: bns_if_source, file: ^bns_if, lets: 0),
+    sugar: bns_route(text: bns_sugar_source, file: ^bns_sugar, lets: 1)
+  }
+}
+
+test fn bare_node_arm_body_before_the_next_arm_takes_the_expression_route() -> Bool {
+  bns_verdicts().arm_before_arm
+}
+
+test fn bare_node_statement_before_an_if_takes_the_expression_route() -> Bool {
+  bns_verdicts().node_before_if
+}
+
+test fn node_let_sugar_still_takes_the_let_route() -> Bool {
+  bns_verdicts().sugar
+}
diff --git a/src/v2/workflow/floor_pure_producer_share.dag b/src/v2/workflow/floor_pure_producer_share.dag
index e6ba567c078..2f49018b5cb 100644
--- a/src/v2/workflow/floor_pure_producer_share.dag
+++ b/src/v2/workflow/floor_pure_producer_share.dag
@@ -637,6 +637,9 @@ import v2.std.collection { List }
 // stores four. Each claim that paid its own parse measured at the new-witness budget's edge on the
 // floor runs of #13056 (eval steps 66-72k, enrolment margin refused on cpu). Portable because each
 // holds only Bools.
+// THE BARE node STATEMENT ROW: v2.test.parse.bare_node_statement_route bns_verdicts parses three
+// one-function modules once each and stores three Bool verdicts, on the same ground. Portable because it
+// holds only Bools.
 // THE dag BIND BLOCK-SCOPED ROW: v2.test.emit.dag_bind_let_block_scoped dbl_verdicts parses five emitted
 // token sequences once each and stores five Bool verdicts; each claim that paid its own parse measured
 // over the new-witness budget on the floor runs of #13099 (eval steps 72-105k). Portable because it
@@ -914,6 +917,7 @@ data floor_cross_claim_pure_producers_warm: List<String> = [
   "v2.test.claim.body_lowering.caret_symbol_value_lowering.csv_verdicts",
   "v2.test.parse.else_less_if_value_category.elif_verdicts",
   "v2.test.parse.brace_and_lambda_head_route.bhr_verdicts",
+  "v2.test.parse.bare_node_statement_route.bns_verdicts",
   "v2.test.emit.dag_bind_let_block_scoped.dbl_verdicts",
   "v2.test.emit.dag_target_text_round_trip.trt_verdicts",
   "v2.test.parse.type_decl_modifier_g0_parse_probe.caret_tree_atom_identities",