Reboot freeze/hang after Grub

OS: Ubuntu Server 16.04.2 / 16.10 / 17.04 (also tried Kernel 4.12.2 in Server 17.04)
Hardware #1:
- Intel NUC6CAYS - Latest BIOS (0040) 6/19/2017
- 2GB Memory CBD16D3LFS1KBG
- ADATA SP550 SSD 120GB
Hardware #2:
- Intel NUC6CAYH - Latest BIOS (0040) 6/19/2017
- 4GB Memory KVR16LS11/4
- PNY CS1311 120GB

I'm building a "kiosk" using Ubuntu Server on Intel NUC. While configuring my system, I noticed that randomly, NUC won't reboot at all. I see GRUB Menu than after the timeout.... nothing (black screen) and I have to power reset. When its freezed, I tried:
* SSH don't work
* CTRL+ALT+F1 / F2 / F3 etc.. don't work

I enabled earlyprintk=efi in grub config, and the hang occurred after bootconsole earlyefi0 disabled:
[ 0.00000] Console: colour dummy device 80x25
[ 0.00000] console [tty0] enabled
[ 0.00000] bootconsole [earlyefi0] disabled
**Freeze here**

## STEPS TO REPRODUCE ##
### INSTALL UBUNTU SERVER ###
Encrypted home: NO
Guided - use entire disk and set up LVM
Force UEFI: YES

### CREATE REBOOT SCRIPTS - EVERY 15 SECONDS ###
tee /usr/local/bin/autoreboot.sh <<'EOF'
#!/bin/bash
sleep 15s
reboot
EOF
chmod +x /usr/local/bin/autoreboot.sh

### CREATE REBOOT SERVICE ###
tee /etc/systemd/system/autoreboot.service <<'EOF'
[Service]
ExecStart=/usr/local/bin/autoreboot.sh

[Install]
WantedBy=default.target
EOF

chmod +x /etc/systemd/system/autoreboot.service
systemctl enable autoreboot.service

### REBOOT AND WAIT ###
reboot

==> When running the script on a NUC with 2GB memory, it will freeze after about 20 reboot (it's really random)
==> When running the script on a NUC with 4GB memory, it will freeze after more than 1 hour, usually about 1h15min.

I can't find any logs of failed boot, or don't know how, since those boots are never logged in journalctl. Hang too early??

Don't know what to try now. Any idea on the cause? Since changing memory have an impact, I ran memtest without issue. I also tried to disable swap but still have the issue

Thanks!