diff --git a/mlir/include/mlir/Dialect/QIR/Transforms/Passes.td b/mlir/include/mlir/Dialect/QIR/Transforms/Passes.td
index d88f5d79fa..74881e5bc1 100644
--- a/mlir/include/mlir/Dialect/QIR/Transforms/Passes.td
+++ b/mlir/include/mlir/Dialect/QIR/Transforms/Passes.td
@@ -16,6 +16,11 @@ def QIRSetAttributesAndMetadata
   let dependentDialects = ["mlir::LLVM::LLVMDialect"];
   let summary = "Sets the required attributes to the entry point function and "
                 "adds the required module flags, compliant with QIR 2.1";
+  let description = [{
+    Static resource counts specify the capacity through the highest qubit ID
+    used by a quantum instruction and the highest recorded result ID. Rejects
+    IDs whose required capacity cannot fit in an unsigned 64-bit integer.
+  }];
   let options = [Option<"useAdaptive", "use-adaptive", "bool",
                         /*default= */ "true", "Specifies the profile.">];
 }
diff --git a/mlir/lib/Dialect/QIR/Transforms/AttachQIRAttributes.cpp b/mlir/lib/Dialect/QIR/Transforms/AttachQIRAttributes.cpp
index 058ee5a8ff..2bf9e09ecf 100644
--- a/mlir/lib/Dialect/QIR/Transforms/AttachQIRAttributes.cpp
+++ b/mlir/lib/Dialect/QIR/Transforms/AttachQIRAttributes.cpp
@@ -27,10 +27,13 @@
 #include <mlir/IR/PatternMatch.h>
 #include <mlir/IR/Value.h>
 #include <mlir/Support/LLVM.h>
+#include <mlir/Support/LogicalResult.h>
 
+#include <algorithm>
 #include <cassert>
 #include <cstddef>
 #include <cstdint>
+#include <limits>
 #include <string>
 #include <tuple>
 #include <utility>
@@ -43,10 +46,10 @@ namespace {
 
 /// State object for tracking QIR metadata during conversion
 struct Metadata {
-  /// Number of qubits used in the module
-  size_t numQubits{0};
-  /// Number of measurement results stored in the module
-  size_t numResults{0};
+  /// Required capacity for static qubit IDs
+  uint64_t numQubits{0};
+  /// Required capacity for static result IDs
+  uint64_t numResults{0};
   /// Whether the module uses dynamic qubit management
   bool useDynamicQubit{false};
   /// Whether the module uses dynamic result management
@@ -81,7 +84,23 @@ struct QIRSetAttributesAndMetadata final
       return;
     }
 
-    Metadata metadata = useAdaptive ? getAdaptive(main) : getBase(main);
+    Metadata metadata = useAdaptive ? getAdaptive(main) : Metadata{};
+    if (!metadata.useDynamicQubit) {
+      const auto numQubits = getNumQubits(main);
+      if (failed(numQubits)) {
+        signalPassFailure();
+        return;
+      }
+      metadata.numQubits = *numQubits;
+    }
+    if (!metadata.useDynamicResult) {
+      const auto numResults = getNumResults(main);
+      if (failed(numResults)) {
+        signalPassFailure();
+        return;
+      }
+      metadata.numResults = *numResults;
+    }
     if (useAdaptive) {
       collectOptionalFeatures(getOperation(), main, metadata);
     }
@@ -96,8 +115,8 @@ struct QIRSetAttributesAndMetadata final
   /// - `entry_point`: Marks the main entry point function
   /// - `output_labeling_schema`: labeled
   /// - `qir_profiles`: base_profile
-  /// - `required_num_qubits`: Number of qubits used
-  /// - `required_num_results`: Number of measurement results
+  /// - `required_num_qubits`: Capacity through the highest static qubit ID
+  /// - `required_num_results`: Capacity through the highest recorded result ID
   /// - `qir_major_version`: 2
   /// - `qir_minor_version`: 1
   /// - `dynamic_qubit_management`: true/false
@@ -228,15 +247,28 @@ struct QIRSetAttributesAndMetadata final
     return preserved;
   }
 
-  /// Count the number of uniquely indexed qubit pointers.
+  /// Extend the resource capacity to include an ID without overflow.
+  static FailureOr<uint64_t> includeResourceId(IntegerAttr index,
+                                               uint64_t capacity,
+                                               Operation* operation) {
+    const auto& value = index.getValue();
+    if (value.getActiveBits() > 64 ||
+        value.getZExtValue() == std::numeric_limits<uint64_t>::max()) {
+      return operation->emitError("static QIR resource ID requires a capacity "
+                                  "that does not fit in 64 bits");
+    }
+    return std::max(capacity, value.getZExtValue() + 1);
+  }
+
+  /// Return the capacity required to address every static qubit ID.
   /// Assumes that qubits are constant integers that are converted to
   /// an integer pointer and then used in (at least) one quantum instruction.
-  static size_t getNumQubits(LLVM::LLVMFuncOp& main) {
+  static FailureOr<uint64_t> getNumQubits(LLVM::LLVMFuncOp& main) {
     static constexpr StringRef QIS_PREFIX = "__quantum__qis";
 
-    DenseSet<APInt> seen;
+    FailureOr<uint64_t> numQubits = uint64_t{0};
     main->walk([&](LLVM::ConstantOp constOp) {
-      if (constOp.use_empty()) {
+      if (failed(numQubits) || constOp.use_empty()) {
         return;
       }
 
@@ -285,18 +317,17 @@ struct QIRSetAttributesAndMetadata final
         return;
       }
 
-      // The set ensures that we don't insert the same index multiple times.
-      seen.insert(intAttr.getValue());
+      numQubits = includeResourceId(intAttr, *numQubits, constOp);
     });
 
-    return seen.size();
+    return numQubits;
   }
 
-  /// Count the number of uniquely indexed result_record_output statements.
-  static size_t getNumResults(LLVM::LLVMFuncOp& main) {
-    DenseSet<APInt> seen;
+  /// Return the capacity required to address every recorded static result ID.
+  static FailureOr<uint64_t> getNumResults(LLVM::LLVMFuncOp& main) {
+    FailureOr<uint64_t> numResults = uint64_t{0};
     main->walk([&](LLVM::CallOp callOp) {
-      if (!callOp.getCallee()) {
+      if (failed(numResults) || !callOp.getCallee()) {
         return;
       }
 
@@ -321,11 +352,10 @@ struct QIRSetAttributesAndMetadata final
         return;
       }
 
-      // The set ensures that we don't insert the same index multiple times.
-      seen.insert(intAttr.getValue());
+      numResults = includeResourceId(intAttr, *numResults, constOp);
     });
 
-    return seen.size();
+    return numResults;
   }
 
   /// Determine whether a loop (as a set of blocks) is an iterative loop (true)
@@ -487,19 +517,7 @@ struct QIRSetAttributesAndMetadata final
     });
   }
 
-  /// Return the metadata for a QIR base profile compliant program.
-  static Metadata getBase(LLVM::LLVMFuncOp& main) {
-    return {
-        .numQubits = getNumQubits(main),
-        .numResults = getNumResults(main),
-        .useDynamicQubit = false,
-        .useDynamicResult = false,
-        .useArrays = false,
-        .backwardsBranching = 0,
-    };
-  }
-
-  /// Return the metadata for a QIR adaptive profile compliant program.
+  /// Return the dynamic resource and control flow metadata for QIR adaptive.
   Metadata getAdaptive(LLVM::LLVMFuncOp& main) {
     const auto& domInfo = getAnalysis<DominanceInfo>();
     const auto [useIteration, useCondTerm] =
@@ -512,14 +530,6 @@ struct QIRSetAttributesAndMetadata final
     md.useDynamicResult = useDynamicResult;
     md.useArrays = useArrays;
 
-    if (!useDynamicQubit) {
-      md.numQubits = getNumQubits(main);
-    }
-
-    if (!useDynamicResult) {
-      md.numResults = getNumResults(main);
-    }
-
     if (useIteration) {
       md.backwardsBranching = useCondTerm ? 3 : 1;
     } else if (useCondTerm) {
diff --git a/mlir/unittests/Compiler/test_compiler_pipeline.cpp b/mlir/unittests/Compiler/test_compiler_pipeline.cpp
index 25c2a3454a..b4d32f2035 100644
--- a/mlir/unittests/Compiler/test_compiler_pipeline.cpp
+++ b/mlir/unittests/Compiler/test_compiler_pipeline.cpp
@@ -50,6 +50,7 @@
 #include <mlir/Dialect/SCF/IR/SCF.h>
 #include <mlir/Dialect/Tensor/IR/Tensor.h>
 #include <mlir/Dialect/UB/IR/UBOps.h>
+#include <mlir/IR/Builders.h>
 #include <mlir/IR/BuiltinAttributes.h>
 #include <mlir/IR/BuiltinOps.h>
 #include <mlir/IR/DialectRegistry.h>
@@ -124,6 +125,8 @@ class CompilerPipelineTest
 protected:
   std::unique_ptr<MLIRContext> context;
 
+  // GoogleTest requires this override name.
+  // NOLINTNEXTLINE(readability-identifier-naming)
   void SetUp() override {
     DialectRegistry registry;
     registry.insert<cbit::CBitDialect, QCDialect, QCODialect,
@@ -1466,6 +1469,45 @@ TEST_F(CompilerPipelineTest, JeffBinaryRoundTripPreservesReusableFunctions) {
   EXPECT_EQ(helper.getNumResults(), 0);
 }
 
+TEST_F(CompilerPipelineTest, QIRPreservesSparseStaticQubitIdsAndCapacity) {
+  auto qc = QCProgram::fromMLIRString(R"mlir(module {
+    func.func @main() attributes {mqt.entry_point} {
+      %q = qc.static 7 : !qc.qubit
+      qc.x %q : !qc.qubit
+      return
+    }
+  })mlir");
+  ASSERT_TRUE(qc);
+  ASSERT_TRUE(succeeded(verify(qc->module())));
+
+  for (const auto format :
+       {ProgramFormat::QIRBase, ProgramFormat::QIRAdaptive}) {
+    auto output = runDefaultPipeline(CompilerInput{qc->copy()}, format);
+    ASSERT_TRUE(output);
+    auto moduleOp = std::get<QIRProgram>(*output).module();
+    ASSERT_TRUE(succeeded(verify(moduleOp)));
+    auto main = getMainFunction(moduleOp);
+    ASSERT_TRUE(main);
+    OpBuilder builder(moduleOp.getContext());
+    EXPECT_TRUE(llvm::is_contained(
+        main.getPassthroughAttr(),
+        builder.getStrArrayAttr({"required_num_qubits", "8"})));
+
+    LLVM::CallOp gate;
+    main.walk([&](LLVM::CallOp call) {
+      if (call.getCallee() == QIR_X) {
+        gate = call;
+      }
+    });
+    ASSERT_TRUE(gate);
+    auto pointer = gate.getOperand(0).getDefiningOp<LLVM::IntToPtrOp>();
+    ASSERT_TRUE(pointer);
+    auto index = pointer.getArg().getDefiningOp<LLVM::ConstantOp>();
+    ASSERT_TRUE(index);
+    EXPECT_EQ(cast<IntegerAttr>(index.getValue()).getInt(), 7);
+  }
+}
+
 TEST_F(CompilerPipelineTest, JeffRejectsMutableClassicalHelperArguments) {
   auto qco = QCOProgram::fromMLIRString(R"mlir(module {
     func.func private @helper(%bits: !cbit.reg<1>) {
diff --git a/mlir/unittests/Dialect/QIR/IR/test_qir_ir.cpp b/mlir/unittests/Dialect/QIR/IR/test_qir_ir.cpp
index f83aed8f45..5775e20a0d 100644
--- a/mlir/unittests/Dialect/QIR/IR/test_qir_ir.cpp
+++ b/mlir/unittests/Dialect/QIR/IR/test_qir_ir.cpp
@@ -19,6 +19,7 @@
 #include <gtest/gtest.h>
 #include <llvm/ADT/APInt.h>
 #include <llvm/ADT/STLExtras.h>
+#include <llvm/ADT/SmallVector.h>
 #include <llvm/IR/Constants.h>
 #include <llvm/IR/LLVMContext.h>
 #include <llvm/IR/Metadata.h>
@@ -36,12 +37,14 @@
 #include <mlir/IR/MLIRContext.h>
 #include <mlir/IR/OperationSupport.h>
 #include <mlir/IR/Verifier.h>
+#include <mlir/Parser/Parser.h>
 #include <mlir/Pass/PassManager.h>
 #include <mlir/Support/LLVM.h>
 
 #include <algorithm>
 #include <array>
 #include <cstddef>
+#include <cstdint>
 #include <iosfwd>
 #include <memory>
 #include <ostream>
@@ -94,6 +97,8 @@ class QIRTest : public testing::TestWithParam<QIRTestCase> {
 protected:
   std::unique_ptr<MLIRContext> context;
 
+  // GoogleTest requires this override name.
+  // NOLINTNEXTLINE(readability-identifier-naming)
   void SetUp() override {
     DialectRegistry registry;
     registry.insert<LLVM::LLVMDialect>();
@@ -309,6 +314,76 @@ TEST_F(QIRTest, PreservesUnrelatedMetadataIdempotently) {
       OperationEquivalence::Flags::None));
 }
 
+TEST_F(QIRTest, MetadataDeclaresCapacityForStaticResourceIds) {
+  struct CapacityCase {
+    SmallVector<int64_t> indices;
+    StringRef requiredCapacity;
+  };
+  const std::array cases{
+      CapacityCase{.indices = {}, .requiredCapacity = "0"},
+      CapacityCase{.indices = {0}, .requiredCapacity = "1"},
+      CapacityCase{.indices = {0, 1, 0}, .requiredCapacity = "2"},
+      CapacityCase{.indices = {7, 2, 7}, .requiredCapacity = "8"},
+  };
+
+  for (const auto profile : {
+           QIRProgramBuilder::Profile::Base,
+           QIRProgramBuilder::Profile::Adaptive,
+       }) {
+    for (const auto& testCase : cases) {
+      SCOPED_TRACE(testing::Message()
+                   << "profile=" << static_cast<int>(profile)
+                   << ", requiredCapacity=" << testCase.requiredCapacity.str());
+      auto moduleOp = QIRProgramBuilder::build(
+          context.get(),
+          [&](QIRProgramBuilder& builder) {
+            for (const auto index : testCase.indices) {
+              auto qubit = builder.staticQubit(index);
+              builder.x(qubit);
+              builder.measure(qubit, index);
+            }
+            return builder.intConstant(0);
+          },
+          profile);
+
+      ASSERT_TRUE(moduleOp);
+      ASSERT_TRUE(succeeded(verify(*moduleOp)));
+      auto main = getMainFunction(moduleOp.get());
+      ASSERT_TRUE(main);
+      const auto passthrough = main.getPassthroughAttr();
+      ASSERT_TRUE(passthrough);
+      OpBuilder builder(context.get());
+      for (const StringRef attribute :
+           {"required_num_qubits", "required_num_results"}) {
+        EXPECT_TRUE(llvm::is_contained(
+            passthrough,
+            builder.getStrArrayAttr({attribute, testCase.requiredCapacity})));
+      }
+    }
+  }
+}
+
+TEST_F(QIRTest, MetadataRejectsUnrepresentableStaticResourceCapacity) {
+  auto moduleOp = parseSourceString<ModuleOp>(R"mlir(module {
+    llvm.func @__quantum__qis__x__body(!llvm.ptr)
+    llvm.func @main() attributes {passthrough = ["entry_point"]} {
+      %index = llvm.mlir.constant(-1 : i64) : i64
+      %qubit = llvm.inttoptr %index : i64 to !llvm.ptr
+      llvm.call @__quantum__qis__x__body(%qubit) : (!llvm.ptr) -> ()
+      llvm.return
+    }
+  })mlir",
+                                              context.get());
+  ASSERT_TRUE(moduleOp);
+  ASSERT_TRUE(succeeded(verify(*moduleOp)));
+  OwningOpRef<ModuleOp> before = moduleOp->clone();
+
+  EXPECT_TRUE(failed(attachQIRMetadata(moduleOp.get())));
+  EXPECT_TRUE(OperationEquivalence::isEquivalentTo(
+      moduleOp.get(), before->getOperation(),
+      OperationEquivalence::Flags::None));
+}
+
 TEST_F(QIRTest, AdaptiveBuilderSelectsControlledSpecializationsByArity) {
   auto module = QIRProgramBuilder::build(
       context.get(),