A single motherboard now ships with a choice between Intel and AMD processors, a configuration DFI says eliminates the need for separate hardware qualification cycles when customers switch CPU vendors. The ARH171 variant runs Intel Core Ultra processors (Series 1), while the ARH173 accepts AMD Ryzen embedded V3000 series chips. Both fit the same Mini-ITX 170mm x 170mm form factor and share peripheral connectors, power requirements, and mounting holes. For system integrators building autonomous mobile robots or collaborative robot cells, the approach means one enclosure design, one cabling harness, and one thermal management system regardless of which processor generation the customer selects.
DFI, a subsidiary of Taiwan-based Qisda Group, announced the platform September 9 at its Taipei headquarters. The company positions the boards as infrastructure for what it calls "physical AI"—machine learning models running on devices that interact with the physical world rather than in data centers. The term has gained traction in robotics circles as transformer-based models move from cloud inference to edge deployment. Both board variants include M.2 slots for AI accelerators, dual 2.5 Gigabit Ethernet ports, and USB4 connectivity. The Intel version supports up to 96GB DDR5 memory; the AMD variant tops out at 64GB. Operating temperature range spans -40°C to 85°C, a specification that matters in warehouse environments where AMRs operate year-round without climate control.
The business case centers on hedging against supply chain volatility and architecture roadmaps that don't always align with product lifecycles. A robotics OEM deploying thousands of units over three years faces risk if a single processor family hits end-of-life or allocation constraints midway through production. Dual-CPU platforms let manufacturers pivot without redesigning carrier boards, redoing FCC certifications, or rewriting low-level firmware. DFI claims the approach cuts time-to-market by six months compared to spinning a new board for each processor generation. The company did not disclose pricing, but Mini-ITX industrial boards in this performance class typically range from $400 to $800 in volume, with AI accelerator modules adding $200 to $1,500 depending on TOPS throughput.
Qisda, DFI's parent, reported $12.3 billion in consolidated revenue for fiscal 2025, with embedded computing products accounting for roughly 18 percent of that total. The company competes with Advantech, Kontron, and AAEON in the industrial motherboard segment, where lead times often stretch 12 to 16 weeks and customers demand seven-year availability guarantees. The ARH171/ARH173 launch follows DFI's April 2026 introduction of the EC51X series for edge AI inference, which used Intel Atom x7000 processors and found adoption in retail checkout kiosks and smart city sensor hubs. The new platform targets a higher performance tier—autonomous forklifts, bin-picking arms, and agricultural robots that run SLAM algorithms and real-time object detection models. DFI says first shipments begin in Q4 2026, with volume production ramping in Q1 2027.
What to Watch: Monitor whether other Tier 1 industrial board makers follow DFI's dual-CPU strategy, particularly Advantech and Kontron, both of which have hinted at flexible processor platforms in recent investor calls. Track adoption among collaborative robot manufacturers like Techman and Hanwha, which typically qualify new compute boards over 90-day test cycles. Watch for DFI's disclosure of design wins at upcoming trade events, especially Automate 2027 in May, where the company typically announces OEM partnerships. Finally, pay attention to Intel and AMD's response—whether they offer commercial incentives or technical support for multi-vendor board designs, or push back by tightening reference design licensing.
