Optical interposer platform for AI datacenter connectivity
POET manufactures custom optical engines and light sources using a proprietary interposer technology designed for AI and datacenter applications. The tech stack reflects a hardware-focused operation: simulation (MATLAB, LabVIEW, JMP), design automation (Jira, Confluence), and statistical process control (Minitab). Active projects cluster heavily around production ramp—NPI implementation, manufacturing flow, process validation, and cleanroom qualification—signaling a transition from R&D into volume manufacturing, while pain points (yield improvement, cost reduction, test automation) confirm early-stage production challenges typical of semiconductor scale-up.
POET Technologies designs and manufactures optical interposer-based chiplets for high-speed networking in AI infrastructure. The company integrates photonic and electronic components into multi-chip modules, reducing assembly complexity and cost versus traditional photonics assembly. Current product focus includes POET Infinity (400G transmit chiplet) and POET Starlight (AI-optimized light engine). Operations span Toronto (HQ), Allentown PA, and Singapore; recent hiring in Malaysia and Singapore suggests Asian manufacturing expansion. With 51–200 employees and steady 6-role open pipeline, the organization is scaling manufacturing and product teams in parallel.
POET manufactures optical engines and light sources using a proprietary Optical Interposer platform that integrates electronic and photonic devices into single multi-chip modules for AI and datacenter applications.
POET is headquartered in Toronto, Ontario, with operations in Allentown, Pennsylvania and Singapore. Recent hiring activity in Malaysia, Singapore, and the United States suggests ongoing geographic expansion.
Other companies in the same industry, closest in size