diff --git a/spark/src/main/scala/io/substrait/spark/expression/ToSparkExpression.scala b/spark/src/main/scala/io/substrait/spark/expression/ToSparkExpression.scala
index 7aaca95bd..749caae6a 100644
--- a/spark/src/main/scala/io/substrait/spark/expression/ToSparkExpression.scala
+++ b/spark/src/main/scala/io/substrait/spark/expression/ToSparkExpression.scala
@@ -148,8 +148,11 @@ class ToSparkExpression(
   override def visit(
       expr: SExpression.IntervalYearLiteral,
       context: EmptyVisitationContext): Literal = {
-    // Spark uses a single months Int as the "physical" type for YearMonthInterval
-    val months = expr.years() * 12 + expr.months()
+    // Spark uses a single months Int as the "physical" type for YearMonthInterval, and both
+    // components are int32, so the flattened total can outrun that carrier: a large year count
+    // used to wrap a positive interval into a negative one. Only the total is significant, so
+    // the check belongs on it rather than on the intermediate product.
+    val months = Math.toIntExact(expr.years() * Util.MONTHS_PER_YEAR + expr.months())
     Literal(months, ToSparkType.convert(expr.getType))
   }
 
diff --git a/spark/src/main/scala/io/substrait/spark/utils/Util.scala b/spark/src/main/scala/io/substrait/spark/utils/Util.scala
index 7ffbe5512..7243a7ca7 100644
--- a/spark/src/main/scala/io/substrait/spark/utils/Util.scala
+++ b/spark/src/main/scala/io/substrait/spark/utils/Util.scala
@@ -23,6 +23,7 @@ object Util {
 
   val SECONDS_PER_DAY: Long = 24 * 60 * 60
   val MICROS_PER_SECOND: Long = 1000 * 1000
+  val MONTHS_PER_YEAR: Long = 12
   val MICROSECOND_PRECISION = 6 // for PrecisionTimestamp(TZ) and IntervalDay types
 
   /** Indexed by exponent, which [[toMicroseconds]] bounds to 0..MICROSECOND_PRECISION. */
diff --git a/spark/src/test/scala/io/substrait/spark/TypesAndLiteralsSuite.scala b/spark/src/test/scala/io/substrait/spark/TypesAndLiteralsSuite.scala
index 8d1b1c491..8a6d2078f 100644
--- a/spark/src/test/scala/io/substrait/spark/TypesAndLiteralsSuite.scala
+++ b/spark/src/test/scala/io/substrait/spark/TypesAndLiteralsSuite.scala
@@ -307,6 +307,28 @@ class TypesAndLiteralsSuite extends SparkFunSuite {
       sparkLiteral(ExpressionCreator.intervalDay(false, Int.MaxValue, 0, 0L, 6)))
   }
 
+  test("a year-month interval reports overflow instead of wrapping") {
+    // years and months are both int32 and Spark's physical type is a months Int, so the flattened
+    // total can outrun the carrier: 178,956,971 years used to come out as -2,147,483,644 months,
+    // and Int.MinValue years as exactly 0 — a zero-length interval. This is the carrier's bound,
+    // not the spec's much tighter 10,000-year one, so it takes a producer already far past that.
+    intercept[ArithmeticException](
+      sparkLiteral(ExpressionCreator.intervalYear(false, 178956971, 0)))
+    intercept[ArithmeticException](
+      sparkLiteral(ExpressionCreator.intervalYear(false, 178956970, 8)))
+    intercept[ArithmeticException](
+      sparkLiteral(ExpressionCreator.intervalYear(false, Int.MinValue, 0)))
+
+    // Only the total is significant, so an intermediate past Int.MaxValue is not itself an error.
+    val mixedSigns = sparkLiteral(ExpressionCreator.intervalYear(false, 178956971, -12))
+    assert(mixedSigns.value === 2147483640)
+
+    // The spec's maximum still converts, and the months component carries through.
+    val atTheBound = sparkLiteral(ExpressionCreator.intervalYear(false, 9999, 12))
+    assert(atTheBound.value === 120000)
+    assert(atTheBound.dataType === YearMonthIntervalType.DEFAULT)
+  }
+
   test("a coarser precision on a type is rejected, since a type has no value to rescale") {
     // A Spark type carries no precision of its own, so mapping precision_timestamp<3> onto
     // TimestampNTZType would reinterpret millisecond counts as microsecond ones. Only the literal