diff --git a/cmd/podman/common/create.go b/cmd/podman/common/create.go
index 921cd5a71e..9ef11ad0c1 100644
--- a/cmd/podman/common/create.go
+++ b/cmd/podman/common/create.go
@@ -56,7 +56,7 @@ func GetCreateFlags(cf *ContainerCLIOpts) *pflag.FlagSet {
 	createFlags.StringVar(
 		&cf.CGroupsMode,
 		"cgroups", containerConfig.Cgroups(),
-		`control container cgroup configuration ("enabled"|"disabled"|"no-conmon")`,
+		`control container cgroup configuration ("enabled"|"disabled"|"no-conmon"|"conmon-delegated")`,
 	)
 	createFlags.StringVar(
 		&cf.CGroupParent,
diff --git a/docs/source/markdown/podman-create.1.md b/docs/source/markdown/podman-create.1.md
index 3ec91a3ad7..289c24d13b 100644
--- a/docs/source/markdown/podman-create.1.md
+++ b/docs/source/markdown/podman-create.1.md
@@ -80,7 +80,7 @@ If the host uses cgroups v1, the default is set to **host**.  On cgroups v2 the
 Determines whether the container will create CGroups.
 Valid values are *enabled*, *disabled*, *no-conmon*, which the default being *enabled*.
 The *disabled* option will force the container to not create CGroups, and thus conflicts with CGroup options (**--cgroupns** and **--cgroup-parent**).
-The *no-conmon* option disables a new CGroup only for the conmon process.
+The *no-conmon* option disables a new CGroup only for the conmon process.  The *conmon-delegated* option reuses the current cgroup for both the conmon and the container payload, it works only on cgroup v2.
 
 **--cgroup-parent**=*path*
 
diff --git a/docs/source/markdown/podman-run.1.md b/docs/source/markdown/podman-run.1.md
index 7e91a06a3e..320a1455b9 100644
--- a/docs/source/markdown/podman-run.1.md
+++ b/docs/source/markdown/podman-run.1.md
@@ -89,7 +89,7 @@ Set the cgroup namespace mode for the container.
 
 If the host uses cgroups v1, the default is set to **host**.  On cgroups v2, the default is **private**.
 
-**--cgroups**=**enabled**|**disabled**|**no-conmon**
+**--cgroups**=**enabled**|**disabled**|**no-conmon**|**conmon-delegated**
 
 Determines whether the container will create CGroups.
 
@@ -97,6 +97,7 @@ Default is **enabled**. The **disabled** option will force the container
 to not create CGroups, and thus conflicts with CGroup options
 (**--cgroupns** and **--cgroup-parent**).
 The **no-conmon** option disables a new CGroup only for the **conmon** process.
+The **conmon-delegated** option reuses the current cgroup for both the conmon and the container payload, it works only on cgroup v2.
 
 **--cgroup-parent**=*path*
 
diff --git a/libpod/container_internal_linux.go b/libpod/container_internal_linux.go
index 12c1abf1c5..1c5ff59889 100644
--- a/libpod/container_internal_linux.go
+++ b/libpod/container_internal_linux.go
@@ -31,6 +31,7 @@ import (
 	"github.com/containers/libpod/pkg/resolvconf"
 	"github.com/containers/libpod/pkg/rootless"
 	"github.com/containers/libpod/pkg/util"
+	"github.com/containers/libpod/utils"
 	"github.com/containers/storage/pkg/archive"
 	securejoin "github.com/cyphar/filepath-securejoin"
 	"github.com/opencontainers/runc/libcontainer/user"
@@ -1505,7 +1506,18 @@ func (c *Container) getOCICgroupPath() (string, error) {
 	case (rootless.IsRootless() && !unified) || c.config.NoCgroups:
 		return "", nil
 	case c.runtime.config.Engine.CgroupManager == config.SystemdCgroupsManager:
-		// When runc is set to use Systemd as a cgroup manager, it
+		if c.config.CgroupsMode == conmonDelegated {
+			if c.config.CgroupParent != "" {
+				return c.config.CgroupParent, nil
+			}
+			selfCgroup, err := utils.GetPidCgroupv2(0)
+			if err != nil {
+				return "", err
+			}
+			return filepath.Join(selfCgroup, "container"), nil
+		}
+
+		// When the OCI runtime is set to use Systemd as a cgroup manager, it
 		// expects cgroups to be passed as follows:
 		// slice:prefix:name
 		systemdCgroups := fmt.Sprintf("%s:libpod:%s", path.Base(c.config.CgroupParent), c.ID())
diff --git a/libpod/oci_conmon.go b/libpod/oci_conmon.go
new file mode 100644
index 0000000000..592852c87e
--- /dev/null
+++ b/libpod/oci_conmon.go
@@ -0,0 +1,7 @@
+package libpod
+
+const (
+	// conmonDelegated is the cgroup mode for reusing the current cgroup both
+	// for conmon and for the container payload.
+	conmonDelegated = "conmon-delegated"
+)
diff --git a/libpod/oci_conmon_linux.go b/libpod/oci_conmon_linux.go
index 625a5bf706..9808f5589f 100644
--- a/libpod/oci_conmon_linux.go
+++ b/libpod/oci_conmon_linux.go
@@ -881,6 +881,12 @@ func (r *ConmonOCIRuntime) createOCIContainer(ctr *Container, restoreOptions *Co
 		return err
 	}
 
+	if ctr.config.CgroupsMode == conmonDelegated {
+		if err := utils.MoveUnderCgroup2Subtree("supervisor"); err != nil {
+			return err
+		}
+	}
+
 	args := r.sharedConmonArgs(ctr, ctr.ID(), ctr.bundlePath(), filepath.Join(ctr.state.RunDir, "pidfile"), ctr.LogPath(), r.exitsDir, ociLog, ctr.LogDriver(), logTag)
 
 	if ctr.config.Spec.Process.Terminal {
@@ -1151,7 +1157,7 @@ func (r *ConmonOCIRuntime) sharedConmonArgs(ctr *Container, cuuid, bundlePath, p
 		"--socket-dir-path", r.socketsDir,
 	}
 
-	if r.cgroupManager == config.SystemdCgroupsManager && !ctr.config.NoCgroups {
+	if r.cgroupManager == config.SystemdCgroupsManager && !ctr.config.NoCgroups && ctr.config.CgroupsMode != conmonDelegated {
 		args = append(args, "-s")
 	}
 
@@ -1253,7 +1259,7 @@ func (r *ConmonOCIRuntime) moveConmonToCgroupAndSignal(ctr *Container, cmd *exec
 
 	// If cgroup creation is disabled - just signal.
 	switch ctr.config.CgroupsMode {
-	case "disabled", "no-conmon":
+	case "disabled", "no-conmon", conmonDelegated:
 		mustCreateCgroup = false
 	}
 
diff --git a/libpod/options.go b/libpod/options.go
index ffc9c10181..cb764024a0 100644
--- a/libpod/options.go
+++ b/libpod/options.go
@@ -1049,7 +1049,7 @@ func WithCgroupsMode(mode string) CtrCreateOption {
 		case "disabled":
 			ctr.config.NoCgroups = true
 			ctr.config.CgroupsMode = mode
-		case "enabled", "no-conmon":
+		case "enabled", "no-conmon", conmonDelegated:
 			ctr.config.CgroupsMode = mode
 		default:
 			return errors.Wrapf(define.ErrInvalidArg, "Invalid cgroup mode %q", mode)
diff --git a/libpod/runtime_ctr.go b/libpod/runtime_ctr.go
index 0431861b53..30c8ca8a5e 100644
--- a/libpod/runtime_ctr.go
+++ b/libpod/runtime_ctr.go
@@ -233,9 +233,9 @@ func (r *Runtime) setupContainer(ctx context.Context, ctr *Container) (_ *Contai
 						return nil, errors.Wrapf(err, "error retrieving pod %s cgroup", pod.ID())
 					}
 					ctr.config.CgroupParent = podCgroup
-				case rootless.IsRootless():
+				case rootless.IsRootless() && ctr.config.CgroupsMode != conmonDelegated:
 					ctr.config.CgroupParent = SystemdDefaultRootlessCgroupParent
-				default:
+				case ctr.config.CgroupsMode != conmonDelegated:
 					ctr.config.CgroupParent = SystemdDefaultCgroupParent
 				}
 			} else if len(ctr.config.CgroupParent) < 6 || !strings.HasSuffix(path.Base(ctr.config.CgroupParent), ".slice") {
diff --git a/utils/utils_supported.go b/utils/utils_supported.go
index ce9fd5604c..f8e49b7527 100644
--- a/utils/utils_supported.go
+++ b/utils/utils_supported.go
@@ -3,10 +3,20 @@
 package utils
 
 import (
+	"bufio"
+	"bytes"
+	"fmt"
+	"io/ioutil"
+	"os"
+	"path/filepath"
+	"strings"
+
 	"github.com/containers/libpod/pkg/cgroups"
 	"github.com/containers/libpod/pkg/rootless"
 	systemdDbus "github.com/coreos/go-systemd/v22/dbus"
 	"github.com/godbus/dbus/v5"
+	"github.com/pkg/errors"
+	"github.com/sirupsen/logrus"
 )
 
 // RunUnderSystemdScope adds the specified pid to a systemd scope
@@ -43,6 +53,77 @@ func RunUnderSystemdScope(pid int, slice string, unitName string) error {
 	return nil
 }
 
+// GetPidCgroupv2 returns the unified cgroup for the specified pid.
+func GetPidCgroupv2(pid int) (string, error) {
+	if pid == 0 {
+		pid = os.Getpid()
+	}
+
+	unified, err := cgroups.IsCgroup2UnifiedMode()
+	if err != nil {
+		return "", err
+	}
+	if !unified {
+		return "", errors.New("move under subtree supported only on cgroup v2")
+	}
+
+	procFile := fmt.Sprintf("/proc/%d/cgroup", pid)
+	f, err := os.Open(procFile)
+	if err != nil {
+		return "", errors.Wrapf(err, "open file %q", procFile)
+	}
+	defer f.Close()
+
+	scanner := bufio.NewScanner(f)
+	cgroup := ""
+	for scanner.Scan() {
+		line := scanner.Text()
+		if strings.HasPrefix(line, "0::") {
+			cgroup = line[3:]
+			break
+		}
+	}
+	if cgroup == "" {
+		return "", errors.Errorf("could not find cgroup v2 mount in %q", procFile)
+	}
+	return cgroup, nil
+
+}
+
+// MoveUnderCgroupSubtree moves the PID under a cgroup subtree.
+func MoveUnderCgroup2Subtree(subtree string) error {
+	cgroup, err := GetPidCgroupv2(0)
+	if err != nil {
+		return err
+	}
+
+	cgroupRoot := "/sys/fs/cgroup"
+
+	processes, err := ioutil.ReadFile(filepath.Join(cgroupRoot, cgroup, "cgroup.procs"))
+	if err != nil {
+		return err
+	}
+
+	newCgroup := filepath.Join(cgroupRoot, cgroup, subtree)
+	if err := os.Mkdir(newCgroup, 0755); err != nil {
+		return err
+	}
+
+	f, err := os.OpenFile(filepath.Join(newCgroup, "cgroup.procs"), os.O_RDWR, 0755)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+
+	for _, pid := range bytes.Split(processes, []byte("\n")) {
+		if _, err := f.Write(pid); err != nil {
+			logrus.Warnf("Cannot move process %s to cgroup %q", pid, newCgroup)
+		}
+	}
+	return nil
+
+}
+
 func newProp(name string, units interface{}) systemdDbus.Property {
 	return systemdDbus.Property{
 		Name:  name,
diff --git a/utils/utils_windows.go b/utils/utils_windows.go
index db27877d96..f92425b894 100644
--- a/utils/utils_windows.go
+++ b/utils/utils_windows.go
@@ -7,3 +7,11 @@ import "github.com/pkg/errors"
 func RunUnderSystemdScope(pid int, slice string, unitName string) error {
 	return errors.New("not implemented for windows")
 }
+
+func MoveUnderCgroup2Subtree(subtree string) error {
+	return errors.New("not implemented for windows")
+}
+
+func GetPidCgroupv2(pid int) (string, error) {
+	return "", errors.New("not implemented for windows")
+}