Your "Obsolete" Phone Is a Better Server Than Your Raspberry Pi

I keep a Motorola Edge 2021 in a drawer. LineageOS 23.2, 8GB RAM, 256GB storage, arm64. Cost me $96 renewed on Amazon. It's currently running Ollama, an AI agent framework, a Discord bot, cron jobs, semantic search across 300+ project files, and still has battery left for actual phone calls. This setup isn't ironic. It's my daily homelab, and it outperforms the Raspberry Pi I almost bought for the same job.

The specs don't lie

| | Motorola Edge 2021 (renewed) | Raspberry Pi 5 (8GB kit) |
|---|---|---|
| Cost | ~$96 | ~$120 + SD card + case + PSU |
| RAM | 8GB LPDDR5 | 8GB LPDDR4X |
| Storage | 256GB UFS (built-in) | microSD or NVMe HAT ($25 extra) |
| Power | USB-C PD, or just... battery | USB-C PD, no battery |
| Network | WiFi 6, or USB-C dock → gigabit ethernet | Gigabit ethernet (built-in) |
| Architecture | arm64 | arm64 |
| OS | LineageOS (current Android, no bloat) | Raspberry Pi OS / Ubuntu |
| UPS | Built-in 5000mAh battery | None, needs external HAT ($40+) |
| Display | 6.7" OLED (optional status monitor) | None, needs HDMI | The phone wins on cost, storage speed, power efficiency, and the fact that it has a built-in UPS that costs extra on a Pi. The Pi's only real advantage is GPIO pins, and I can count on one hand the times I've actually needed those in a server context.

What I'm actually running

Here's what's running right now:
  • Termux with full build toolchain: clang, rust, cargo, nodejs

  • Ollama serving kimi-k2.6:cloud, glm-5.2, qwen2.5 locally

  • Hermes Agent in proot Ubuntu with deadsnakes Python 3.13

  • OpenClaw gateway with cron, subagents, multi-model routing

  • MemPalace semantic search across 318 project drawers

  • Discord bot via discord.py

  • Custom scripts: tclaw (OpenClaw maintenance), compass (project manager), mms-fetch (Android MMS extraction)
    • All of this runs on a device that fits in my pocket, draws less power than a lightbulb, and survives power outages for hours.

    The USB-C dock is the secret weapon

    A $25 USB-C dock with PD passthrough and ethernet turns this from "phone with a terminal app" into "actual headless server." Plug in power, ethernet, set the display to never sleep, and you've got a silent, low-power, always-on compute node. The battery acts as a built-in UPS. Power goes out? Phone stays up. Mine made it through three outages this year without a hiccup.

    Why this matters

    We throw away hundreds of millions of phones every year. Most have years of useful life left. LineageOS keeps the software current. The hardware is faster than most IoT boards. And the total cost to repurpose one as a server is under $125 including the dock. The "customization rabbit hole" is where the real learning happens. Every workaround teaches you something that "it just works" never would. Getting maturin to compile Rust wheels on Android Bionic. Setting up deadsnakes PPA in proot Ubuntu. Wiring up a Discord bot that lives inside a phone. These aren't hacks. They're infrastructure skills that transfer directly to cloud and server environments.

    The Pi isn't dead, but it's not the default anymore

    I still have a Pi. It runs Home Assistant and a few sensors. But for general compute — AI inference, agent frameworks, development environments — the phone has become my go-to. The only thing I miss is the GPIO headers, and for that I just use a $5 ESP32. If you're looking for cheap, silent, low-power compute, check your junk drawer first. That old phone might surprise you. --- Running on: Motorola Edge 2021, LineageOS 23.2, Termux, proot Ubuntu with deadsnakes Python 3.13, Ollama, and enough caffeine to be dangerous. Uploaded Image

    Comments

    Post a Comment

    Popular posts from this blog

    Variational Reasoning and Metastability

    Image
      Navigating the AI Landscape: How I Learned to Explore 'Thought Gravity' Have you ever felt like you’re having a conversation with an AI, and suddenly, the AI just... gets stuck? Not stuck on a fact, but stuck on a perspective . It’s like trying to steer a conversation out of a deep canyon; the walls are steep, and every path leads back to the same well-trodden ground.   I’ve been spending a lot of time presenting my AI collaborators with complex problems, and I’ve noticed something fascinating. Our ideas, our questions, and even our creativity don’t exist in a flat, open field. They exist in a landscape defined by "gravity wells. " Think of it this way: an AI’s inference engine is like an incredibly sophisticated, context-aware spell-checker. It’s always looking for the most probable path . So, if an idea has been well-represented in its training data—if millions of people have written about it, and the internet is saturated with examples—that ide...
    Read more

    SMS Session Connector: AI in Your Pocket, No Cloud Required

    Image
    I built something wild this week. My phone now has a persistent AI agent you can text — and it texts back. No Twilio. No cloud webhooks. Just a rooted Android, Termux, and some bash. The Stack 800 lines of bash running as a daemon on a Motorola Edge termux-api for native SMS send/receive (no third-party SMS gateways) OpenClaw for the AI brain — persistent session with full conversation memory whisper.cpp running locally for voice memo transcription ffmpeg for AMR→WAV conversion on inbound MMS audio runit for service supervision (auto-restarts, survives reboots) How It Works 1. You text the number. The daemon polls termux-sms-list every 5 seconds. 2. New message? Forwarded to an OpenClaw persistent session via cron agent turn. 3. The AI processes it — full access to skills, tools, web search, memory. 4. Reply gets written to an outbox file, daemon sends it back via termux-sms-send . 5. Voice memos? MMS audio → mms-fetch → ffmpeg → whisper.cpp → transcript → same pipe...
    Read more

    Multiple LLM Providers, the joy

    Image
      So at the moment, I am running OpenClaw running Ollama. In addition, I am using Gemini to create images, such as this one, and Claude to fix things when they break. That is the joyful part. The not so joyful part is that I am running into limits of the Google Api for doing ordinary things, such as setting a reminder in tasks, or creating an email with an inline image. When things break, and the frequently do, I need to use Claude Code to fix things. It keeps me entertained.
    Read more