diff --git a/.bazelrc b/.bazelrc
index 2b7bf08..6bbff86 100644
--- a/.bazelrc
+++ b/.bazelrc
@@ -1,14 +1,6 @@
 # Always tell why tests fail
 test --test_output=errors
 
-# The performance gains offered by the lockfile are minimal for us.
-# Pinning the dependencies should be superfluous from a reproducibility perspective as the central registry is supposed
-# to be immutable for a module X in version Y after it has ben registered with the registry (ignoring the possibility of
-# yanking a module). Furthermore, the module resolution is deterministic.
-# However, there are several open issues suggesting the lock file will change in one way or another.
-# Thus, we don't use the bzlmod locking yet.
-common --lockfile_mode=off
-
 # When working with hermetic Python toolchains, supporting the legacy runfiles layout is needlessly wasting resources.
 # See https://github.com/bazelbuild/rules_python/issues/1653
 common --nolegacy_external_runfiles
diff --git a/.bazelversion b/.bazelversion
index 21c8c7b..b26a34e 100644
--- a/.bazelversion
+++ b/.bazelversion
@@ -1 +1 @@
-7.1.1
+7.2.1
diff --git a/.gitignore b/.gitignore
index fd501c7..6cdb89b 100644
--- a/.gitignore
+++ b/.gitignore
@@ -2,6 +2,7 @@
 
 # Bazel
 bazel-*
+*MODULE.bazel.lock
 user.bazelrc
 
 # Python
diff --git a/examples/.bazelrc b/examples/.bazelrc
index 96b86b7..085b611 100644
--- a/examples/.bazelrc
+++ b/examples/.bazelrc
@@ -8,11 +8,3 @@ common --nolegacy_external_runfiles
 
 # The symlinks are annoying
 common --experimental_convenience_symlinks=clean
-
-# The performance gains offered by the lockfile are minimal for us.
-# Pinning the dependencies should be superfluous from a reproducibility perspective as the central registry is supposed
-# to be immutable for a module X in version Y after it has ben registered with the registry (ignoring the possibility of
-# yanking a module). Furthermore, the module resolution is deterministic.
-# However, there are several open issues suggesting the lock file will change in one way or another.
-# Thus, we don't use the bzlmod locking yet.
-common --lockfile_mode=off
diff --git a/examples/.bazelversion b/examples/.bazelversion
index 66ce77b..b26a34e 100644
--- a/examples/.bazelversion
+++ b/examples/.bazelversion
@@ -1 +1 @@
-7.0.0
+7.2.1
diff --git a/test/apply_fixes/execution_logic.py b/test/apply_fixes/execution_logic.py
index 0ea458b..2cc2602 100644
--- a/test/apply_fixes/execution_logic.py
+++ b/test/apply_fixes/execution_logic.py
@@ -42,10 +42,11 @@
 # Decrease Python toolchain overhead
 # See https://github.com/bazelbuild/rules_python/issues/1653
 common --nolegacy_external_runfiles
+# Useless as the workspace are thrown away after each creation
 common --lockfile_mode=off
 """
 
-BAZEL_VERSION = "7.0.0"
+BAZEL_VERSION = "7.2.1"
 
 
 def dwyu_path_as_string(dwyu_path: Path) -> str:
diff --git a/test/aspect/.bazelrc b/test/aspect/.bazelrc
index 21a3131..0c01517 100644
--- a/test/aspect/.bazelrc
+++ b/test/aspect/.bazelrc
@@ -4,11 +4,3 @@ common --nolegacy_external_runfiles
 
 # The symlinks are annoying
 common --experimental_convenience_symlinks=clean
-
-# The performance gains offered by the lockfile are minimal for us.
-# Pinning the dependencies should be superfluous from a reproducibility perspective as the central registry is supposed
-# to be immutable for a module X in version Y after it has ben registered with the registry (ignoring the possibility of
-# yanking a module). Furthermore, the module resolution is deterministic.
-# However, there are several open issues suggesting the lock file will change in one way or another.
-# Thus, we don't use the bzlmod locking yet.
-common --lockfile_mode=off
diff --git a/test/aspect/execute_tests.py b/test/aspect/execute_tests.py
index 657ffdc..788e4fc 100755
--- a/test/aspect/execute_tests.py
+++ b/test/aspect/execute_tests.py
@@ -25,11 +25,6 @@
     # We support Bazel's modern dependency management system, but it works only as desired with a recent Bazel version
     "--experimental_enable_bzlmod=false": CompatibleVersions(before="6.0.0"),
     "--enable_bzlmod=false": CompatibleVersions(minimum="6.0.0", before="6.2.0"),
-    # We are not yet sure if we really want to lock the bzlmod resolution down given we test with various Bazel versions
-    # and configurations. It seems the main benefits of the lock file are not having to reanalyze the central registry
-    # when working without a cached workspace and being safeguarded against changed or yanked modules in the central
-    # registry. Both don't matter much to us right now.
-    "--lockfile_mode=off": CompatibleVersions(minimum="6.2.0"),
     # Incompatible changes
     "--incompatible_legacy_local_fallback=false": CompatibleVersions(minimum="5.0.0"),  # false is the forward path
     "--incompatible_enforce_config_setting_visibility": CompatibleVersions(minimum="5.0.0"),