Why

Every current Dockerfile targets x86_64 only. As ARM64 hardware proliferates the install matrix needs to follow... NVIDIA DGX Spark (GB10 Grace + Blackwell, ARM CPU + Blackwell GPU, shipping Q1-Q2 2026), Snapdragon X Elite laptops (Adreno GPU + Hexagon NPU), Apple Silicon containers, Jetson Orin/Thor, Raspberry Pi 5, AWS Graviton... none of these can even pull our images today.

Pairs with PC-114 (#114). Vulkan covers the GPU diversity axis, ARM64 covers the CPU arch axis... together they unlock basically every modern compute target outside the macOS-native install path (#32).

What changes

1. .github/workflows/build-images.yml gets docker buildx + QEMU setup, builds for linux/amd64,linux/arm64, pushes multi-arch GHCR manifests. Users keep pulling the same tag and get the right arch automatically.

2. Per-Dockerfile arm64 audit:

   • Dockerfile.v31 (CUDA): NVIDIA ships nvidia/cuda:*-ubuntu22.04 arm64 variants for CUDA 12.x (GB10 is on this line). Need to verify the exact tag we pin has arm64 + that the patched llama.cpp build works on aarch64 CUDA.
   • Dockerfile.rocm: AMD doesn't ship arm64 ROCm at all currently. Skip arm64 for this image until/unless that changes.
   • Dockerfile.vulkan (feat: Vulkan universal backend (one image, covers NVIDIA + AMD + Intel + Apple + CPU) #114): ubuntu:22.04 + mesa-vulkan-drivers, both arm64-native. Should Just Work but needs validation.
   • proxy/Dockerfile: Go binary, just GOOS=linux GOARCH=arm64 in the buildx target.
   • geometric-lens/Dockerfile, v3-service/Dockerfile, sandbox/Dockerfile: Python + apt, all arm64-native on Ubuntu.

3. tier.py: detect_gpu() is mostly arch-agnostic but the nvidia-smi/rocm-smi output parsers might trip on the GB10's Grace+Blackwell topology... verify on actual hardware when available.

4. atlas doctor: docker-pull tests should add --platform when a specific service has only amd64 (so users on arm64 hosts get a clear 'no arm64 build for X' message instead of pulling the wrong arch and crashing).

5. Docs: SETUP.md gets an arm64 matrix table (which Dockerfile supports which arch per release).

What this unlocks

Hardware testing matrix

Each combo needs a tester...

• DGX Spark + CUDA: waiting on hardware (NVIDIA Q1-Q2 2026)
• Apple Silicon + Vulkan-in-Docker: need a Mac dev
• Snapdragon X Elite + Vulkan: need a Snapdragon laptop owner
• Jetson Orin + CUDA: any Jetson user willing
• Pi 5 + Vulkan: cheap, can probably pick one up

Out of scope

• Per-arch performance tuning (one ticket per arch if perf is bad)
• Apple Silicon NATIVE install path (feat: Apple Silicon (Metal) deployment guide #32) is the no-Docker fast path, tracked separately
• Windows ARM64 is a separate Docker-on-Windows story

Dependencies

• PC-114 (feat: Vulkan universal backend (one image, covers NVIDIA + AMD + Intel + Apple + CPU) #114) for the Vulkan universal case to work on arm64 hosts that don't have CUDA/ROCm available.