Silicom Secures Initial AI Inference Orders with Multi-Thousand-Dollar ASP Potential
Silicom secured initial orders for its inference-specific solution to power a Proof of Concept in H2 2026, with first-half 2026 deliveries and potential full-scale deployment of tens of thousands of units at a multi-thousand-dollar ASP. The company has also won orders from two AI compute contenders for separate inference products.
1. AI Inference PoC Engagement
Silicom received initial orders for its inference-specific acceleration solution to support a Proof of Concept targeting Tier-1 hyperscalers in the second half of 2026. Deliveries are scheduled for the first half of 2026 and, pending PoC success, could scale to tens of thousands of units with a multi-thousand-dollar average selling price each.
2. Growing Product Traction
The company has secured orders from two leading AI compute challengers for distinct inference products and is engaged in developing a third inference-specific solution. This builds on Silicom’s adaptable FPGA and COTS networking architecture designed to reduce latency and total cost of ownership for large-scale AI inference deployments.
3. Needham Conference Participation
A senior Silicom executive will present at the Needham Technology, Media & Consumer Conference on May 12, 2026, in New York. The presentation aims to highlight Silicom’s AI inference strategy and advance discussions with institutional investors through one-on-one meetings.