Configurations
6 specced configs — ABM-enrolled, r1o pre-loaded, ready to ship.
Showing reference catalog. Live inventory updates when backend connects.
Services & Add-ons
Don't just buy hardware — get a production-ready cluster. Pre-loaded models, RDMA mesh, distributed inference, automated workflows. Ships configured.
Unit Setup
Single-unit r1o stack install, configured and burn-tested.
- r1o stack pre-loaded (Tailscale, monitoring, inference serving)
- macOS hardened + energy settings optimized
- 72-hour burn-in (GPU stress, memory, thermal, I/O)
- Apple Business enrollment via Configurator
Model Pre-load
Ship with inference-ready models already downloaded and verified.
- Up to 3 MLX models downloaded and quantized
- Model serving configured (mlx-lm or vLLM-MLX)
- Inference endpoint verified before shipping
- Model config persisted across reboots
Cluster Build
Full multi-unit cluster: RDMA mesh, distributed inference, monitoring.
- Thunderbolt 5 RDMA mesh configured
- JACCL distributed inference across all units
- Unified monitoring dashboard (asmi)
- Model sharding and load balancing
- Tailscale mesh + SSH key distribution
- Full cluster health verification
Enterprise Cluster
5+ units with custom topology, on-site setup, and dedicated support.
- Everything in Cluster Build
- Custom network topology design
- On-site installation available
- 90-day post-deploy support
- Automated failover and recovery
- Dedicated Slack/Teams channel
Workflow Automation
Automated pipelines: model serving, health checks, alerts, scaling.
- Automated model deployment pipeline
- Health monitoring with iMessage/Slack alerts
- Auto-restart on failure with recovery agents
- Scheduled model updates and quantization
- API gateway with load balancing
Thunderbolt 5 Cables
Certified TB5 cables for RDMA mesh. 120 Gb/s, 240W charging. Included with cluster builds, available separately for single units.
- 0.8m and 2m lengths available
- 120 Gb/s bi-directional bandwidth
- 240W power delivery
- Required for RDMA mesh between units
r1o Agent Stack
Full AI ops agent pre-installed. ABM enrollment, device management, model serving, cluster health — all managed by an autonomous agent with MCP tools.
- Hermes Agent with 30+ operational skills
- Apple Business MCP — device enrollment, MDM, compliance
- Cluster intelligence (asmi) — GPU/RAM/power/RDMA monitoring
- Model inventory management across all units
- Automated sourcing pipeline — price watching, deal alerts
- iMessage/Slack notifications for health events
- Self-healing recovery agents for inference failures
- NanoMDM integration — push profiles, manage fleet