diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -271,6 +271,11 @@
 def SVLD1SW : MInst<"svld1sw_{d}", "dPU", "lUl",             [IsLoad],               MemEltTyInt32,   "aarch64_sve_ld1">;
 def SVLD1UW : MInst<"svld1uw_{d}", "dPY", "lUl",             [IsLoad, IsZExtReturn], MemEltTyInt32,   "aarch64_sve_ld1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVLD1_BF      : MInst<"svld1[_{2}]",      "dPc",  "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ld1">;
+  def SVLD1_VNUM_BF : MInst<"svld1_vnum[_{2}]", "dPcl", "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ld1">;
+}
+
 // Load one vector (scalar base, VL displacement)
 def SVLD1_VNUM   : MInst<"svld1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfd", [IsLoad],               MemEltTyDefault, "aarch64_sve_ld1">;
 def SVLD1SB_VNUM : MInst<"svld1sb_vnum_{d}", "dPSl", "silUsUiUl",       [IsLoad],               MemEltTyInt8,    "aarch64_sve_ld1">;
@@ -376,6 +381,11 @@
 def SVLDFF1SW_VNUM : MInst<"svldff1sw_vnum_{d}", "dPUl", "lUl",             [IsLoad],               MemEltTyInt32,   "aarch64_sve_ldff1">;
 def SVLDFF1UW_VNUM : MInst<"svldff1uw_vnum_{d}", "dPYl", "lUl",             [IsLoad, IsZExtReturn], MemEltTyInt32,   "aarch64_sve_ldff1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVLDFF1_BF      : MInst<"svldff1[_{2}]",      "dPc",  "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldff1">;
+  def SVLDFF1_VNUM_BF : MInst<"svldff1_vnum[_{2}]", "dPcl", "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldff1">;
+}
+
 // First-faulting load one vector (vector base)
 def SVLDFF1_GATHER_BASES_U   : MInst<"svldff1_gather[_{2}base]_{d}",   "dPu", "ilUiUlfd", [IsGatherLoad],               MemEltTyDefault, "aarch64_sve_ldff1_gather_scalar_offset">;
 def SVLDFF1SB_GATHER_BASES_U : MInst<"svldff1sb_gather[_{2}base]_{d}", "dPu", "ilUiUl",   [IsGatherLoad],               MemEltTyInt8,    "aarch64_sve_ldff1_gather_scalar_offset">;
@@ -471,15 +481,29 @@
 def SVLDNF1SW_VNUM : MInst<"svldnf1sw_vnum_{d}", "dPUl", "lUl",             [IsLoad],               MemEltTyInt32,   "aarch64_sve_ldnf1">;
 def SVLDNF1UW_VNUM : MInst<"svldnf1uw_vnum_{d}", "dPYl", "lUl",             [IsLoad, IsZExtReturn], MemEltTyInt32,   "aarch64_sve_ldnf1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVLDNF1_BF      : MInst<"svldnf1[_{2}]",      "dPc",  "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnf1">;
+  def SVLDNF1_VNUM_BF : MInst<"svldnf1_vnum[_{2}]", "dPcl", "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnf1">;
+}
+
 // Load one vector, unextended load, non-temporal (scalar base)
 def SVLDNT1 : MInst<"svldnt1[_{2}]", "dPc", "csilUcUsUiUlhfd", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnt1">;
 
 // Load one vector, unextended load, non-temporal (scalar base, VL displacement)
 def SVLDNT1_VNUM : MInst<"svldnt1_vnum[_{2}]", "dPcl", "csilUcUsUiUlhfd", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnt1">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVLDNT1_BF      : MInst<"svldnt1[_{2}]",      "dPc",  "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnt1">;
+  def SVLDNT1_VNUM_BF : MInst<"svldnt1_vnum[_{2}]", "dPcl", "b", [IsLoad], MemEltTyDefault, "aarch64_sve_ldnt1">;
+}
+
 // Load one quadword and replicate (scalar base)
 def SVLD1RQ : SInst<"svld1rq[_{2}]", "dPc", "csilUcUsUiUlhfd", MergeNone, "aarch64_sve_ld1rq">;
 
+let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
+  def SVLD1RQ_BF : SInst<"svld1rq[_{2}]", "dPc",  "b", MergeNone, "aarch64_sve_ld1rq">;
+}
+
 multiclass StructLoad<string name, string proto, string i> {
   def : SInst<name, proto, "csilUcUsUiUlhfd", MergeNone, i, [IsStructLoad]>;
   let ArchGuard = "defined(__ARM_FEATURE_SVE_BF16)" in {
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1-bfloat.c
@@ -0,0 +1,34 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
+
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svbfloat16_t test_svld1_bf16(svbool_t pg, const bfloat16_t *base)
+{
+  // CHECK-LABEL: test_svld1_bf16
+  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svld1_bf16'}}
+  return SVE_ACLE_FUNC(svld1,_bf16,,)(pg, base);
+}
+
+svbfloat16_t test_svld1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
+{
+  // CHECK-LABEL: test_svld1_vnum_bf16
+  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
+  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svld1_vnum_bf16'}}
+  return SVE_ACLE_FUNC(svld1_vnum,_bf16,,)(pg, base, vnum);
+}
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ld1rq-bfloat.c
@@ -0,0 +1,22 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
+
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svbfloat16_t test_svld1rq_bf16(svbool_t pg, const bfloat16_t *base)
+{
+  // CHECK-LABEL: test_svld1rq_bf16
+  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[INTRINSIC:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
+  // CHECK: ret <vscale x 8 x bfloat> %[[INTRINSIC]]
+  // expected-warning@+1 {{implicit declaration of function 'svld1rq_bf16'}}
+  return SVE_ACLE_FUNC(svld1rq,_bf16,,)(pg, base);
+}
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldff1-bfloat.c
@@ -0,0 +1,34 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
+
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svbfloat16_t test_svldff1_bf16(svbool_t pg, const bfloat16_t *base)
+{
+  // CHECK-LABEL: test_svldff1_bf16
+  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldff1_bf16'}}
+  return SVE_ACLE_FUNC(svldff1,_bf16,,)(pg, base);
+}
+
+svbfloat16_t test_svldff1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
+{
+  // CHECK-LABEL: test_svldff1_vnum_bf16
+  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
+  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
+  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldff1_vnum_bf16'}}
+  return SVE_ACLE_FUNC(svldff1_vnum,_bf16,,)(pg, base, vnum);
+}
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnf1-bfloat.c
@@ -0,0 +1,33 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svbfloat16_t test_svldnf1_bf16(svbool_t pg, const bfloat16_t *base)
+{
+  // CHECK-LABEL: test_svldnf1_bf16
+  // CHECK: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldnf1_bf16'}}
+  return SVE_ACLE_FUNC(svldnf1,_bf16,,)(pg, base);
+}
+
+svbfloat16_t test_svldnf1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
+{
+  // CHECK-LABEL: test_svldnf1_vnum_bf16
+  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
+  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
+  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldnf1_vnum_bf16'}}
+  return SVE_ACLE_FUNC(svldnf1_vnum,_bf16,,)(pg, base, vnum);
+}
diff --git a/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_ldnt1-bfloat.c
@@ -0,0 +1,34 @@
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -D__ARM_FEATURE_SVE_BF16 -DSVE_OVERLOADED_FORMS -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -S -O1 -Werror -Wall -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -D__ARM_FEATURE_SVE -D__ARM_FEATURE_BF16_SCALAR_ARITHMETIC -triple aarch64-none-linux-gnu -target-feature +sve -target-feature +bf16 -fallow-half-arguments-and-returns -fsyntax-only -verify -verify-ignore-unexpected=error -verify-ignore-unexpected=note %s
+
+#include <arm_sve.h>
+
+#ifdef SVE_OVERLOADED_FORMS
+// A simple used,unused... macro, long enough to represent any SVE builtin.
+#define SVE_ACLE_FUNC(A1,A2_UNUSED,A3,A4_UNUSED) A1##A3
+#else
+#define SVE_ACLE_FUNC(A1,A2,A3,A4) A1##A2##A3##A4
+#endif
+
+svbfloat16_t test_svldnt1_bf16(svbool_t pg, const bfloat16_t *base)
+{
+  // CHECK-LABEL: test_svldnt1_bf16
+  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %base)
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldnt1_bf16'}}
+  return SVE_ACLE_FUNC(svldnt1,_bf16,,)(pg, base);
+}
+
+svbfloat16_t test_svldnt1_vnum_bf16(svbool_t pg, const bfloat16_t *base, int64_t vnum)
+{
+  // CHECK-LABEL: test_svldnt1_vnum_bf16
+  // CHECK-DAG: %[[PG:.*]] = call <vscale x 8 x i1> @llvm.aarch64.sve.convert.from.svbool.nxv8i1(<vscale x 16 x i1> %pg)
+  // CHECK-DAG: %[[BITCAST:.*]] = bitcast bfloat* %base to <vscale x 8 x bfloat>*
+  // CHECK-DAG: %[[GEP:.*]] = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %[[BITCAST]], i64 %vnum, i64 0
+  // CHECK: %[[LOAD:.*]] = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> %[[PG]], bfloat* %[[GEP]])
+  // CHECK: ret <vscale x 8 x bfloat> %[[LOAD]]
+  // expected-warning@+1 {{implicit declaration of function 'svldnt1_vnum_bf16'}}
+  return SVE_ACLE_FUNC(svldnt1_vnum,_bf16,,)(pg, base, vnum);
+}
diff --git a/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td b/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
--- a/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
+++ b/llvm/lib/Target/AArch64/AArch64SVEInstrInfo.td
@@ -1518,10 +1518,11 @@
   defm : pred_load<nxv4f32, nxv4i1, nonext_masked_load,    LD1W,    LD1W_IMM,    am_sve_regreg_lsl2>;
 
   // 8-element contiguous loads
-  defm : pred_load<nxv8i16, nxv8i1, zext_masked_load_i8,  LD1B_H,  LD1B_H_IMM,  am_sve_regreg_lsl0>;
-  defm : pred_load<nxv8i16, nxv8i1, asext_masked_load_i8, LD1SB_H, LD1SB_H_IMM, am_sve_regreg_lsl0>;
-  defm : pred_load<nxv8i16, nxv8i1, nonext_masked_load,   LD1H,    LD1H_IMM,    am_sve_regreg_lsl1>;
-  defm : pred_load<nxv8f16, nxv8i1, nonext_masked_load,   LD1H,    LD1H_IMM,    am_sve_regreg_lsl1>;
+  defm : pred_load<nxv8i16,  nxv8i1, zext_masked_load_i8,  LD1B_H,  LD1B_H_IMM,  am_sve_regreg_lsl0>;
+  defm : pred_load<nxv8i16,  nxv8i1, asext_masked_load_i8, LD1SB_H, LD1SB_H_IMM, am_sve_regreg_lsl0>;
+  defm : pred_load<nxv8i16,  nxv8i1, nonext_masked_load,   LD1H,    LD1H_IMM,    am_sve_regreg_lsl1>;
+  defm : pred_load<nxv8f16,  nxv8i1, nonext_masked_load,   LD1H,    LD1H_IMM,    am_sve_regreg_lsl1>;
+  defm : pred_load<nxv8bf16, nxv8i1, nonext_masked_load,   LD1H,    LD1H_IMM,    am_sve_regreg_lsl1>;
 
   // 16-element contiguous loads
   defm : pred_load<nxv16i8, nxv16i1, nonext_masked_load, LD1B, LD1B_IMM, am_sve_regreg_lsl0>;
@@ -1704,10 +1705,11 @@
   defm : ld1<LD1W,    LD1W_IMM,    nxv4f32, AArch64ld1,  nxv4i1, nxv4f32, am_sve_regreg_lsl2>;
 
   // 8-element contiguous loads
-  defm : ld1<LD1B_H,  LD1B_H_IMM,  nxv8i16, AArch64ld1,  nxv8i1, nxv8i8,  am_sve_regreg_lsl0>;
-  defm : ld1<LD1SB_H, LD1SB_H_IMM, nxv8i16, AArch64ld1s, nxv8i1, nxv8i8,  am_sve_regreg_lsl0>;
-  defm : ld1<LD1H,    LD1H_IMM,    nxv8i16, AArch64ld1,  nxv8i1, nxv8i16, am_sve_regreg_lsl1>;
-  defm : ld1<LD1H,    LD1H_IMM,    nxv8f16, AArch64ld1,  nxv8i1, nxv8f16, am_sve_regreg_lsl1>;
+  defm : ld1<LD1B_H,  LD1B_H_IMM,  nxv8i16,  AArch64ld1,  nxv8i1, nxv8i8,   am_sve_regreg_lsl0>;
+  defm : ld1<LD1SB_H, LD1SB_H_IMM, nxv8i16,  AArch64ld1s, nxv8i1, nxv8i8,   am_sve_regreg_lsl0>;
+  defm : ld1<LD1H,    LD1H_IMM,    nxv8i16,  AArch64ld1,  nxv8i1, nxv8i16,  am_sve_regreg_lsl1>;
+  defm : ld1<LD1H,    LD1H_IMM,    nxv8f16,  AArch64ld1,  nxv8i1, nxv8f16,  am_sve_regreg_lsl1>;
+  defm : ld1<LD1H,    LD1H_IMM,    nxv8bf16, AArch64ld1,  nxv8i1, nxv8bf16, am_sve_regreg_lsl1>;
 
   // 16-element contiguous loads
   defm : ld1<LD1B, LD1B_IMM, nxv16i8, AArch64ld1, nxv16i1, nxv16i8, am_sve_regreg_lsl0>;
@@ -1725,31 +1727,32 @@
   }
 
   // 2-element contiguous non-faulting loads
-  defm : ldnf1<LDNF1B_D_IMM,  nxv2i64, AArch64ldnf1,  nxv2i1, nxv2i8>;
-  defm : ldnf1<LDNF1SB_D_IMM, nxv2i64, AArch64ldnf1s, nxv2i1, nxv2i8>;
-  defm : ldnf1<LDNF1H_D_IMM,  nxv2i64, AArch64ldnf1,  nxv2i1, nxv2i16>;
-  defm : ldnf1<LDNF1SH_D_IMM, nxv2i64, AArch64ldnf1s, nxv2i1, nxv2i16>;
-  defm : ldnf1<LDNF1W_D_IMM,  nxv2i64, AArch64ldnf1,  nxv2i1, nxv2i32>;
-  defm : ldnf1<LDNF1SW_D_IMM, nxv2i64, AArch64ldnf1s, nxv2i1, nxv2i32>;
-  defm : ldnf1<LDNF1D_IMM,    nxv2i64, AArch64ldnf1,  nxv2i1, nxv2i64>;
-  defm : ldnf1<LDNF1D_IMM,    nxv2f64, AArch64ldnf1,  nxv2i1, nxv2f64>;
+  defm : ldnf1<LDNF1B_D_IMM,  nxv2i64,  AArch64ldnf1,  nxv2i1, nxv2i8>;
+  defm : ldnf1<LDNF1SB_D_IMM, nxv2i64,  AArch64ldnf1s, nxv2i1, nxv2i8>;
+  defm : ldnf1<LDNF1H_D_IMM,  nxv2i64,  AArch64ldnf1,  nxv2i1, nxv2i16>;
+  defm : ldnf1<LDNF1SH_D_IMM, nxv2i64,  AArch64ldnf1s, nxv2i1, nxv2i16>;
+  defm : ldnf1<LDNF1W_D_IMM,  nxv2i64,  AArch64ldnf1,  nxv2i1, nxv2i32>;
+  defm : ldnf1<LDNF1SW_D_IMM, nxv2i64,  AArch64ldnf1s, nxv2i1, nxv2i32>;
+  defm : ldnf1<LDNF1D_IMM,    nxv2i64,  AArch64ldnf1,  nxv2i1, nxv2i64>;
+  defm : ldnf1<LDNF1D_IMM,    nxv2f64,  AArch64ldnf1,  nxv2i1, nxv2f64>;
 
   // 4-element contiguous non-faulting loads
-  defm : ldnf1<LDNF1B_S_IMM,  nxv4i32, AArch64ldnf1,  nxv4i1, nxv4i8>;
-  defm : ldnf1<LDNF1SB_S_IMM, nxv4i32, AArch64ldnf1s, nxv4i1, nxv4i8>;
-  defm : ldnf1<LDNF1H_S_IMM,  nxv4i32, AArch64ldnf1,  nxv4i1, nxv4i16>;
-  defm : ldnf1<LDNF1SH_S_IMM, nxv4i32, AArch64ldnf1s, nxv4i1, nxv4i16>;
-  defm : ldnf1<LDNF1W_IMM,    nxv4i32, AArch64ldnf1,  nxv4i1, nxv4i32>;
-  defm : ldnf1<LDNF1W_IMM,    nxv4f32, AArch64ldnf1,  nxv4i1, nxv4f32>;
+  defm : ldnf1<LDNF1B_S_IMM,  nxv4i32,  AArch64ldnf1,  nxv4i1, nxv4i8>;
+  defm : ldnf1<LDNF1SB_S_IMM, nxv4i32,  AArch64ldnf1s, nxv4i1, nxv4i8>;
+  defm : ldnf1<LDNF1H_S_IMM,  nxv4i32,  AArch64ldnf1,  nxv4i1, nxv4i16>;
+  defm : ldnf1<LDNF1SH_S_IMM, nxv4i32,  AArch64ldnf1s, nxv4i1, nxv4i16>;
+  defm : ldnf1<LDNF1W_IMM,    nxv4i32,  AArch64ldnf1,  nxv4i1, nxv4i32>;
+  defm : ldnf1<LDNF1W_IMM,    nxv4f32,  AArch64ldnf1,  nxv4i1, nxv4f32>;
 
   // 8-element contiguous non-faulting loads
-  defm : ldnf1<LDNF1B_H_IMM,  nxv8i16, AArch64ldnf1,  nxv8i1, nxv8i8>;
-  defm : ldnf1<LDNF1SB_H_IMM, nxv8i16, AArch64ldnf1s, nxv8i1, nxv8i8>;
-  defm : ldnf1<LDNF1H_IMM,    nxv8i16, AArch64ldnf1,  nxv8i1, nxv8i16>;
-  defm : ldnf1<LDNF1H_IMM,    nxv8f16, AArch64ldnf1,  nxv8i1, nxv8f16>;
+  defm : ldnf1<LDNF1B_H_IMM,  nxv8i16,  AArch64ldnf1,  nxv8i1, nxv8i8>;
+  defm : ldnf1<LDNF1SB_H_IMM, nxv8i16,  AArch64ldnf1s, nxv8i1, nxv8i8>;
+  defm : ldnf1<LDNF1H_IMM,    nxv8i16,  AArch64ldnf1,  nxv8i1, nxv8i16>;
+  defm : ldnf1<LDNF1H_IMM,    nxv8f16,  AArch64ldnf1,  nxv8i1, nxv8f16>;
+  defm : ldnf1<LDNF1H_IMM,    nxv8bf16, AArch64ldnf1,  nxv8i1, nxv8bf16>;
 
   // 16-element contiguous non-faulting loads
-  defm : ldnf1<LDNF1B_IMM,    nxv16i8, AArch64ldnf1, nxv16i1, nxv16i8>;
+  defm : ldnf1<LDNF1B_IMM,    nxv16i8,  AArch64ldnf1, nxv16i1, nxv16i8>;
 
   multiclass ldff1<Instruction I, ValueType Ty, SDPatternOperator Load, ValueType PredTy, ValueType MemVT, ComplexPattern AddrCP> {
     // reg + reg
@@ -1764,29 +1767,30 @@
   }
 
   // 2-element contiguous first faulting loads
-  defm : ldff1<LDFF1B_D,  nxv2i64, AArch64ldff1,  nxv2i1, nxv2i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1SB_D, nxv2i64, AArch64ldff1s, nxv2i1, nxv2i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1H_D,  nxv2i64, AArch64ldff1,  nxv2i1, nxv2i16, am_sve_regreg_lsl1>;
-  defm : ldff1<LDFF1SH_D, nxv2i64, AArch64ldff1s, nxv2i1, nxv2i16, am_sve_regreg_lsl1>;
-  defm : ldff1<LDFF1W_D,  nxv2i64, AArch64ldff1,  nxv2i1, nxv2i32, am_sve_regreg_lsl2>;
-  defm : ldff1<LDFF1SW_D, nxv2i64, AArch64ldff1s, nxv2i1, nxv2i32, am_sve_regreg_lsl2>;
-  defm : ldff1<LDFF1D,    nxv2i64, AArch64ldff1,  nxv2i1, nxv2i64, am_sve_regreg_lsl3>;
-  defm : ldff1<LDFF1W_D,  nxv2f32, AArch64ldff1,  nxv2i1, nxv2f32, am_sve_regreg_lsl2>;
-  defm : ldff1<LDFF1D,    nxv2f64, AArch64ldff1,  nxv2i1, nxv2f64, am_sve_regreg_lsl3>;
+  defm : ldff1<LDFF1B_D,  nxv2i64,  AArch64ldff1,  nxv2i1, nxv2i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1SB_D, nxv2i64,  AArch64ldff1s, nxv2i1, nxv2i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1H_D,  nxv2i64,  AArch64ldff1,  nxv2i1, nxv2i16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1SH_D, nxv2i64,  AArch64ldff1s, nxv2i1, nxv2i16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1W_D,  nxv2i64,  AArch64ldff1,  nxv2i1, nxv2i32,  am_sve_regreg_lsl2>;
+  defm : ldff1<LDFF1SW_D, nxv2i64,  AArch64ldff1s, nxv2i1, nxv2i32,  am_sve_regreg_lsl2>;
+  defm : ldff1<LDFF1D,    nxv2i64,  AArch64ldff1,  nxv2i1, nxv2i64,  am_sve_regreg_lsl3>;
+  defm : ldff1<LDFF1W_D,  nxv2f32,  AArch64ldff1,  nxv2i1, nxv2f32,  am_sve_regreg_lsl2>;
+  defm : ldff1<LDFF1D,    nxv2f64,  AArch64ldff1,  nxv2i1, nxv2f64,  am_sve_regreg_lsl3>;
 
   // 4-element contiguous first faulting loads
-  defm : ldff1<LDFF1B_S,  nxv4i32, AArch64ldff1,  nxv4i1, nxv4i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1SB_S, nxv4i32, AArch64ldff1s, nxv4i1, nxv4i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1H_S,  nxv4i32, AArch64ldff1,  nxv4i1, nxv4i16, am_sve_regreg_lsl1>;
-  defm : ldff1<LDFF1SH_S, nxv4i32, AArch64ldff1s, nxv4i1, nxv4i16, am_sve_regreg_lsl1>;
-  defm : ldff1<LDFF1W,    nxv4i32, AArch64ldff1,  nxv4i1, nxv4i32, am_sve_regreg_lsl2>;
-  defm : ldff1<LDFF1W,    nxv4f32, AArch64ldff1,  nxv4i1, nxv4f32, am_sve_regreg_lsl2>;
+  defm : ldff1<LDFF1B_S,  nxv4i32,  AArch64ldff1,  nxv4i1, nxv4i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1SB_S, nxv4i32,  AArch64ldff1s, nxv4i1, nxv4i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1H_S,  nxv4i32,  AArch64ldff1,  nxv4i1, nxv4i16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1SH_S, nxv4i32,  AArch64ldff1s, nxv4i1, nxv4i16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1W,    nxv4i32,  AArch64ldff1,  nxv4i1, nxv4i32,  am_sve_regreg_lsl2>;
+  defm : ldff1<LDFF1W,    nxv4f32,  AArch64ldff1,  nxv4i1, nxv4f32,  am_sve_regreg_lsl2>;
 
   // 8-element contiguous first faulting loads
-  defm : ldff1<LDFF1B_H,  nxv8i16, AArch64ldff1,  nxv8i1, nxv8i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1SB_H, nxv8i16, AArch64ldff1s, nxv8i1, nxv8i8,  am_sve_regreg_lsl0>;
-  defm : ldff1<LDFF1H,    nxv8i16, AArch64ldff1,  nxv8i1, nxv8i16, am_sve_regreg_lsl1>;
-  defm : ldff1<LDFF1H,    nxv8f16, AArch64ldff1,  nxv8i1, nxv8f16, am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1B_H,  nxv8i16,  AArch64ldff1,  nxv8i1, nxv8i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1SB_H, nxv8i16,  AArch64ldff1s, nxv8i1, nxv8i8,   am_sve_regreg_lsl0>;
+  defm : ldff1<LDFF1H,    nxv8i16,  AArch64ldff1,  nxv8i1, nxv8i16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1H,    nxv8f16,  AArch64ldff1,  nxv8i1, nxv8f16,  am_sve_regreg_lsl1>;
+  defm : ldff1<LDFF1H,    nxv8bf16, AArch64ldff1,  nxv8i1, nxv8bf16, am_sve_regreg_lsl1>;
 
   // 16-element contiguous first faulting loads
   defm : ldff1<LDFF1B, nxv16i8, AArch64ldff1, nxv16i1, nxv16i8, am_sve_regreg_lsl0>;
diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h
@@ -161,7 +161,8 @@
       return false;
 
     Type *Ty = cast<VectorType>(DataType)->getElementType();
-    if (Ty->isHalfTy() || Ty->isFloatTy() || Ty->isDoubleTy())
+    if (Ty->isBFloatTy() || Ty->isHalfTy() ||
+        Ty->isFloatTy() || Ty->isDoubleTy())
       return true;
 
     if (Ty->isIntegerTy(8) || Ty->isIntegerTy(16) ||
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-imm.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-imm.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-imm.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-imm.ll
@@ -207,6 +207,17 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ld1h_bf16_inbound(<vscale x 8 x i1> %pg, bfloat* %a) {
+; CHECK-LABEL: ld1h_bf16_inbound:
+; CHECK: ld1h { z0.h }, p0/z, [x0, #1, mul vl]
+; CHECK-NEXT: ret
+  %base_scalable = bitcast bfloat* %a to <vscale x 8 x bfloat>*
+  %base = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %base_scalable, i64 1
+  %base_scalar = bitcast <vscale x 8 x bfloat>* %base to bfloat*
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %base_scalar)
+  ret <vscale x 8 x bfloat> %load
+}
+
 ;
 ; LD1W
 ;
@@ -288,6 +299,7 @@
 declare <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1>, i8*)
 declare <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1>, i16*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 
 declare <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1>, i8*)
 declare <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1>, i16*)
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-reg.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-reg.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-reg.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1-addressing-mode-reg-reg.ll
@@ -95,6 +95,15 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ld1h_bf16(<vscale x 8 x i1> %pg, bfloat* %a, i64 %index) {
+; CHECK-LABEL: ld1h_bf16
+; CHECK: ld1h { z0.h }, p0/z, [x0, x1, lsl #1]
+; CHECK-NEXT: ret
+  %base = getelementptr bfloat, bfloat* %a, i64 %index
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %base)
+  ret <vscale x 8 x bfloat> %load
+}
+
 define <vscale x 4 x i32> @ld1h_s(<vscale x 4 x i1> %pred, i16* %a, i64 %index) {
 ; CHECK-LABEL: ld1h_s:
 ; CHECK: ld1h { z0.s }, p0/z, [x0, x1, lsl #1]
@@ -204,6 +213,7 @@
 declare <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1>, i8*)
 declare <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1>, i16*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 
 declare <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1>, i8*)
 declare <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1>, i16*)
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-ld1.ll
@@ -87,6 +87,14 @@
   ret <vscale x 8 x half> %res
 }
 
+define <vscale x 8 x bfloat> @ld1h_bf16(<vscale x 8 x i1> %pred, bfloat* %addr) {
+; CHECK-LABEL: ld1h_bf16:
+; CHECK: ld1h { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1> %pred, bfloat* %addr)
+  ret <vscale x 8 x bfloat> %res
+}
+
 define <vscale x 4 x i32> @ld1h_s(<vscale x 4 x i1> %pred, i16* %addr) {
 ; CHECK-LABEL: ld1h_s:
 ; CHECK: ld1h { z0.s }, p0/z, [x0]
@@ -188,6 +196,7 @@
 declare <vscale x 8 x i8> @llvm.aarch64.sve.ld1.nxv8i8(<vscale x 8 x i1>, i8*)
 declare <vscale x 8 x i16> @llvm.aarch64.sve.ld1.nxv8i16(<vscale x 8 x i1>, i16*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ld1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 
 declare <vscale x 4 x i8> @llvm.aarch64.sve.ld1.nxv4i8(<vscale x 4 x i1>, i8*)
 declare <vscale x 4 x i16> @llvm.aarch64.sve.ld1.nxv4i16(<vscale x 4 x i1>, i16*)
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-ff.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-ff.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-ff.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-ff.ll
@@ -206,6 +206,14 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ldff1h_bf16(<vscale x 8 x i1> %pg, bfloat* %a) {
+; CHECK-LABEL: ldff1h_bf16:
+; CHECK: ldff1h { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %a)
+  ret <vscale x 8 x bfloat> %load
+}
+
 define <vscale x 8 x half> @ldff1h_f16_reg(<vscale x 8 x i1> %pg, half* %a, i64 %offset) {
 ; CHECK-LABEL: ldff1h_f16_reg:
 ; CHECK: ldff1h { z0.h }, p0/z, [x0, x1, lsl #1]
@@ -215,6 +223,15 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ldff1h_bf16_reg(<vscale x 8 x i1> %pg, bfloat* %a, i64 %offset) {
+; CHECK-LABEL: ldff1h_bf16_reg:
+; CHECK: ldff1h { z0.h }, p0/z, [x0, x1, lsl #1]
+; CHECK-NEXT: ret
+  %base = getelementptr bfloat, bfloat* %a, i64 %offset
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %base)
+  ret <vscale x 8 x bfloat> %load
+}
+
 ;
 ; LDFF1SH
 ;
@@ -398,6 +415,7 @@
 declare <vscale x 8 x i8> @llvm.aarch64.sve.ldff1.nxv8i8(<vscale x 8 x i1>, i8*)
 declare <vscale x 8 x i16> @llvm.aarch64.sve.ldff1.nxv8i16(<vscale x 8 x i1>, i16*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ldff1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ldff1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 
 declare <vscale x 4 x i8> @llvm.aarch64.sve.ldff1.nxv4i8(<vscale x 4 x i1>, i8*)
 declare <vscale x 4 x i16> @llvm.aarch64.sve.ldff1.nxv4i16(<vscale x 4 x i1>, i16*)
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-nf.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-nf.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-nf.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads-nf.ll
@@ -140,6 +140,14 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ldnf1h_bf16(<vscale x 8 x i1> %pg, bfloat* %a) {
+; CHECK-LABEL: ldnf1h_bf16:
+; CHECK: ldnf1h { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %a)
+  ret <vscale x 8 x bfloat> %load
+}
+
 define <vscale x 8 x half> @ldnf1h_f16_inbound(<vscale x 8 x i1> %pg, half* %a) {
 ; CHECK-LABEL: ldnf1h_f16_inbound:
 ; CHECK: ldnf1h { z0.h }, p0/z, [x0, #1, mul vl]
@@ -151,6 +159,17 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @ldnf1h_bf16_inbound(<vscale x 8 x i1> %pg, bfloat* %a) {
+; CHECK-LABEL: ldnf1h_bf16_inbound:
+; CHECK: ldnf1h { z0.h }, p0/z, [x0, #1, mul vl]
+; CHECK-NEXT: ret
+  %base_scalable = bitcast bfloat* %a to <vscale x 8 x bfloat>*
+  %base = getelementptr <vscale x 8 x bfloat>, <vscale x 8 x bfloat>* %base_scalable, i64 1
+  %base_scalar = bitcast <vscale x 8 x bfloat>* %base to bfloat*
+  %load = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1> %pg, bfloat* %base_scalar)
+  ret <vscale x 8 x bfloat> %load
+}
+
 define <vscale x 4 x i32> @ldnf1b_s(<vscale x 4 x i1> %pg, i8* %a) {
 ; CHECK-LABEL: ldnf1b_s:
 ; CHECK: ldnf1b { z0.s }, p0/z, [x0]
@@ -442,6 +461,7 @@
 declare <vscale x 8 x i8> @llvm.aarch64.sve.ldnf1.nxv8i8(<vscale x 8 x i1>, i8*)
 declare <vscale x 8 x i16> @llvm.aarch64.sve.ldnf1.nxv8i16(<vscale x 8 x i1>, i16*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ldnf1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnf1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 
 declare <vscale x 4 x i8> @llvm.aarch64.sve.ldnf1.nxv4i8(<vscale x 4 x i1>, i8*)
 declare <vscale x 4 x i16> @llvm.aarch64.sve.ldnf1.nxv4i16(<vscale x 4 x i1>, i16*)
diff --git a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads.ll b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads.ll
--- a/llvm/test/CodeGen/AArch64/sve-intrinsics-loads.ll
+++ b/llvm/test/CodeGen/AArch64/sve-intrinsics-loads.ll
@@ -97,6 +97,23 @@
   ret <vscale x 8 x half> %res
 }
 
+define <vscale x 8 x bfloat> @ld1rqh_bf16(<vscale x 8 x i1> %pred, bfloat* %addr) {
+; CHECK-LABEL: ld1rqh_bf16:
+; CHECK: ld1rqh { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> %pred, bfloat* %addr)
+  ret <vscale x 8 x bfloat> %res
+}
+
+define <vscale x 8 x bfloat> @ld1rqh_bf16_imm(<vscale x 8 x i1> %pred, bfloat* %addr) {
+; CHECK-LABEL: ld1rqh_bf16_imm:
+; CHECK: ld1rqh { z0.h }, p0/z, [x0, #-16]
+; CHECK-NEXT: ret
+  %ptr = getelementptr inbounds bfloat, bfloat* %addr, i16 -8
+  %res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1> %pred, bfloat* %ptr)
+  ret <vscale x 8 x bfloat> %res
+}
+
 ;
 ; LD1RQW
 ;
@@ -208,6 +225,15 @@
   ret <vscale x 8 x half> %res
 }
 
+define <vscale x 8 x bfloat> @ldnt1h_bf16(<vscale x 8 x i1> %pred, bfloat* %addr) {
+; CHECK-LABEL: ldnt1h_bf16:
+; CHECK: ldnt1h { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %res = call <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1> %pred,
+                                                                     bfloat* %addr)
+  ret <vscale x 8 x bfloat> %res
+}
+
 ;
 ; LDNT1W
 ;
@@ -498,6 +524,7 @@
 declare <vscale x 4 x i32> @llvm.aarch64.sve.ld1rq.nxv4i32(<vscale x 4 x i1>, i32*)
 declare <vscale x 2 x i64> @llvm.aarch64.sve.ld1rq.nxv2i64(<vscale x 2 x i1>, i64*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ld1rq.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ld1rq.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 declare <vscale x 4 x float> @llvm.aarch64.sve.ld1rq.nxv4f32(<vscale x 4 x i1>, float*)
 declare <vscale x 2 x double> @llvm.aarch64.sve.ld1rq.nxv2f64(<vscale x 2 x i1>, double*)
 
@@ -506,6 +533,7 @@
 declare <vscale x 4 x i32> @llvm.aarch64.sve.ldnt1.nxv4i32(<vscale x 4 x i1>, i32*)
 declare <vscale x 2 x i64> @llvm.aarch64.sve.ldnt1.nxv2i64(<vscale x 2 x i1>, i64*)
 declare <vscale x 8 x half> @llvm.aarch64.sve.ldnt1.nxv8f16(<vscale x 8 x i1>, half*)
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.ldnt1.nxv8bf16(<vscale x 8 x i1>, bfloat*)
 declare <vscale x 4 x float> @llvm.aarch64.sve.ldnt1.nxv4f32(<vscale x 4 x i1>, float*)
 declare <vscale x 2 x double> @llvm.aarch64.sve.ldnt1.nxv2f64(<vscale x 2 x i1>, double*)
 
diff --git a/llvm/test/CodeGen/AArch64/sve-masked-ldst-nonext.ll b/llvm/test/CodeGen/AArch64/sve-masked-ldst-nonext.ll
--- a/llvm/test/CodeGen/AArch64/sve-masked-ldst-nonext.ll
+++ b/llvm/test/CodeGen/AArch64/sve-masked-ldst-nonext.ll
@@ -87,6 +87,14 @@
   ret <vscale x 8 x half> %load
 }
 
+define <vscale x 8 x bfloat> @masked_load_nxv8bf16(<vscale x 8 x bfloat> *%a, <vscale x 8 x i1> %mask) nounwind {
+; CHECK-LABEL: masked_load_nxv8bf16:
+; CHECK-NEXT: ld1h { z0.h }, p0/z, [x0]
+; CHECK-NEXT: ret
+  %load = call <vscale x 8 x bfloat> @llvm.masked.load.nxv8bf16(<vscale x 8 x bfloat> *%a, i32 2, <vscale x 8 x i1> %mask, <vscale x 8 x bfloat> undef)
+  ret <vscale x 8 x bfloat> %load
+}
+
 ;
 ; Masked Stores
 ;
@@ -182,6 +190,7 @@
 declare <vscale x 4 x float> @llvm.masked.load.nxv4f32(<vscale x 4 x float>*, i32, <vscale x 4 x i1>, <vscale x 4 x float>)
 declare <vscale x 4 x half> @llvm.masked.load.nxv4f16(<vscale x 4 x half>*, i32, <vscale x 4 x i1>, <vscale x 4 x half>)
 declare <vscale x 8 x half> @llvm.masked.load.nxv8f16(<vscale x 8 x half>*, i32, <vscale x 8 x i1>, <vscale x 8 x half>)
+declare <vscale x 8 x bfloat> @llvm.masked.load.nxv8bf16(<vscale x 8 x bfloat>*, i32, <vscale x 8 x i1>, <vscale x 8 x bfloat>)
 
 declare void @llvm.masked.store.nxv2i64(<vscale x 2 x i64>, <vscale x 2 x i64>*, i32, <vscale x 2 x i1>)
 declare void @llvm.masked.store.nxv4i32(<vscale x 4 x i32>, <vscale x 4 x i32>*, i32, <vscale x 4 x i1>)