Hardware Architecture Overview
SAM: The Private AI Appliance
Peripheral Connectivity
Tripp Lite Video Adapter + HDMI / USB Cabling
SAM Core Intelligence Engine
- Runs local AI orchestration (Guardian)
- Processes internal automation workflows
- Hosts private SAM Intelligence services
SAM Expansion Base
- Additional USB Ports
- SD Card Storage Expansion
- Peripheral Access Layer
Fiber-Bandwidth Connectivity
Dedicated high-bandwidth tunnel for zero-latency data processing between intelligence and local storage.
Data Intelligence Hub
- Primary AI data storage
- Encrypted Knowledge Base
- Local Vector / Dataset store
Recommended Standards
Primary BackupEvery SAM includes a high-speed 4 TB backup drive for local data redundancy.
Off-Campus RedundancyWe recommend a second 4 TB backup drive that is rotated weekly and taken off-campus for physical security.
The Local Deputy Strategy
SAM is designed to keep your AI intelligence "inside the building" on your own hardware. The strength of that promise scales with the hardware tier you choose.
By processing data locally, we reduce cloud surveillance exposure and keep your competitive intelligence under your control. For enterprise deployments, local SAM Deputies integrate with the NVIDIA NemoClaw Sheriff for corporate-wide perimeter security.
SAM Hardware Tiers
Three Appliances. One Product Family.
SAM ships in three hardware configurations, each paired with a service tier. The hardware you choose determines how much of SAM's intelligence runs locally versus through secure cloud assistance — and it determines which service tier you're eligible to subscribe to.
One rule to know before you buy
Hardware is the ceiling. Service tier is what you subscribe to inside that ceiling. You can always pair down — Premium hardware can run Essentials service if that's all you need. But you cannot pair up. Starter hardware cannot host Professional service. Standard hardware cannot host Enterprise service. If your needs grow beyond your hardware tier, the hardware is physically swapped, not upgraded in software.
SAM Essentials
A dedicated AI assistant for everyday work. Private memory, conversational local intelligence, cloud-assisted task execution.
- Telegram executive assistant (always available, 24/7)
- Email triage and drafting
- Web research and summarization
- Continuous memory across every conversation
- Monthly maintenance by Motherboard technicians
Local memory, private conversations. Tool execution routes through encrypted cloud APIs because 16 GB of unified memory cannot reliably host the larger models required for on-device tool operations.
SAM Professional
Local-first AI with reliable on-device tool execution. Everything in Essentials, plus deep integrations with the tools you already use.
- Everything in Essentials
- CRM integration (GoHighLevel, HubSpot, Salesforce — choose one)
- Accounting integration (QuickBooks Online or Xero — choose one)
- Microsoft 365, Slack, Zoom meeting intelligence
- LinkedIn automation and prospecting
- Browser automation for web portals and dashboards
- Document creation — reports, proposals, briefings
Local memory, private conversations, local tool execution for most workflows. 24 GB of unified memory hosts a capable on-device model that handles your daily business tasks without reaching out to cloud services. Cloud escalation is reserved for heavy reasoning tasks that exceed local capacity — and even those remain governed by your cost-sensitivity settings.
SAM Enterprise
Fully local intelligence for operators managing complexity at scale. Multi-entity support, unlimited integrations, proactive monitoring.
- Everything in Professional
- Multi-entity support (multiple businesses under one SAM)
- Unlimited custom integrations
- Investor analysis and board preparation
- Proactive monitoring and scheduled intelligence reports
- Weekly optimization review with dedicated Motherboard team
- ToggleLogic.ai licensing path for building your own AI products
Fully local intelligence. Private memory, private conversations, local tool execution across the full stack. 48 GB or more of unified memory hosts the community-validated best-in-class on-device models for agentic work. Cloud is touched only for explicitly-approved strategic work where you have judged the value warrants it.
Enterprise Cluster Deployments
Beyond a Single Appliance
For enterprise corporations managing AI across departments, SAM scales from clustered Mac mini deployments to full NVIDIA DGX integration with the NemoClaw Sheriff.
Mac Mini Cluster (SMB Enterprise)
Ideal for: Small to mid-sized businesses with 5–50 employees needing departmental AI
- Departmental AI deployment (HR, Sales, Support, IT)
- Each department operates its own SAM Deputy
- Cross-department data isolation by design
- Silent operation for office environments
- Apple ecosystem integration
NVIDIA DGX SuperPOD (Corporate Enterprise)
Ideal for: Enterprise corporations with 500+ employees
- NVIDIA NemoClaw Sheriff integration
- Multi-departmental AI orchestration at scale
- High-performance GPU acceleration
- Enterprise-grade redundancy & uptime
- Datacenter deployment architecture
The Hybrid Strategy: Sheriff + Deputy at Scale
For corporate enterprise deployments, SAM runs as a Sovereign Deputy in each department (HR, Legal, Sales, IT) while the NVIDIA NemoClaw Sheriff provides the global security perimeter.
This architecture delivers:
- Departmental Privacy: Each Deputy operates in isolation with zero cross-department data leakage
- Corporate Governance: The Sheriff enforces company-wide AI policies and security standards
- Scalable Intelligence: Add new departmental Deputies without disrupting existing operations
- Local Processing: Corporate intelligence stays within corporate infrastructure
Not sure which tier fits? We'll help you figure it out.