FabuBlox has introduced FabAgent to help engineers prepare semiconductor fabrication flows before cleanroom work. The assistant combines process instructions, device-layer visualization and knowledge of a facility’s tools and material restrictions. Engineers retain review of proposed steps. The company is connecting design history with physical wafer tracking, while more autonomous factories remain a longer-term development goal.
FabuBlox, a semiconductor-software company spun out of the Massachusetts Institute of Technology in the US, introduced FabAgent to help engineers prepare fabrication flows. The AI assistant connects proposed steps to knowledge of the facility that will execute them. Engineers can build and inspect a process before starting cleanroom work, retaining review of the proposed design.[1]
The platform’s Facility OS stores equipment capabilities, material restrictions, parameter ranges and recipes. Those details bring the tools actually available into process planning. A visualization follows how successive fabrication steps change the device’s layers. Recipes stay linked to their equipment, and engineers can choose which recipe changes to carry into a process flow as the design develops.[1]
Fab Run Studio, the platform’s wafer-tracking component, is in early beta and connects process versions with physical wafers, execution history and measurements. It covers branched routes and experiments using different parameters. The company’s longer-term plans include specialized agents learning from physical outcomes and helping coordinate experiments. Live equipment signals and increasingly autonomous fabrication remain development directions around the current engineering assistant.[1]
