A VPN does use battery, but on a modern phone with a light protocol it is a small addition, not "half the battery." Power goes to two things: encrypting every packet and keeping the tunnel alive in the background. WireGuard and VLESS do both cheaply; OpenVPN is noticeably worse. Most of the drain that phone statistics attribute to the VPN actually belongs to the apps whose traffic passes through it.

Where the power goes

  • Encryption. Modern ciphers (ChaCha20, AES with hardware acceleration) handle gigabytes with almost no CPU. On phones from the last several years this is a fraction of a percent.
  • Keeping the tunnel alive. To keep the connection from being dropped behind the carrier's NAT, the client sends small keepalive packets periodically, each briefly waking the radio. WireGuard does this every 25 seconds and is silent when there is no data; OpenVPN maintains more state and talks more often.
  • Reconnections. Every network change means a new handshake. WireGuard's is a single cheap exchange; OpenVPN's is heavy. On a train or with frequent Wi-Fi and mobile switches, the difference adds up.
  • The radio. The biggest consumer is not encryption but transmission. The VPN adds a small header to each packet, so the radio is on slightly longer; that is the real cost.

Why the statistics mislead

Android and iOS attribute data and battery to the app that initiated the traffic. With a VPN on, everything passes through the VPN app, and the system charges all of it to the VPN: video, photo sync, updates. So the VPN "uses 20%" when without it the same 20% would have been charged to YouTube and the cloud. To see the real cost, compare total drain across a day with and without the VPN at similar usage, not the VPN line in the statistics.

How to measure

  1. A day without the VPN, normal use: battery in the morning and evening, screen-on time.
  2. A day with the VPN, similar use.
  3. The difference in percentage points at similar screen time is the real cost. If it is large, look for a cause below.

What increases the drain

  • OpenVPN, especially in TCP mode.
  • A poor connection: lost packets are resent, the radio works longer. Not the VPN's fault, but visible.
  • Frequent reconnections from unstable Wi-Fi or aggressive battery management that kills the tunnel and then rebuilds it.
  • A distant server: more retransmissions and waiting.
  • Multi-hop: two encryptions instead of one.

What reduces it

  1. WireGuard or VLESS instead of OpenVPN.
  2. A nearby, unloaded server: fewer losses, fewer retries; latency is on the server list.
  3. Unrestricted in battery settings for the VPN app. It sounds backwards, but the restriction is what causes the kill-and-reconnect cycle that costs more than a quiet tunnel. The Android setup is in how to set up a VPN on Android.
  4. Split tunneling: heavy background traffic that does not need privacy outside the tunnel.
  5. Updates on Wi-Fi and while charging: not about the VPN, but the main background consumer.
  6. Wi-Fi over mobile data where you have the choice: the Wi-Fi radio is cheaper.

Should you turn it off to save power?

Not on unfamiliar networks: the saving is minimal and the protection is gone entirely. At home on your own Wi-Fi it is your call. A sensible compromise is auto-connect on untrusted networks and manual control at home. What a VPN does and does not protect on a phone is in what is a VPN.

How much data, rather than battery, a VPN uses is covered separately in does a VPN use more data.

404 VPN uses WireGuard and VLESS, the most efficient protocols, tunes keepalive for mobile networks, and picks the server by latency automatically. Get started here.