SEMIFIVE has signed a turnkey contract with Mobilint to develop an AI semiconductor for robotics applications, the companies announced. The project falls under South Korea's "K-On-Device AI Semiconductor" program, a government-backed initiative aimed at strengthening the domestic edge-AI chip sector.

Under the agreement, SEMIFIVE will handle chip development on a turnkey basis for Mobilint, a Korean AI chip startup focused on edge inference hardware. The target device is an on-device AI accelerator designed for robotics workloads, where computation runs locally on the machine rather than in the cloud.

The K-On-Device AI Semiconductor program, administered by the Korean government, selects and funds local companies to build processors for on-device AI applications. The initiative forms part of a broader national push to reduce reliance on foreign chip suppliers and build domestic capabilities in edge AI silicon.

Turnkey contracts of this type transfer the burden of design implementation, verification and chip bring-up to the design service provider. SEMIFIVE, a design solution partner specializing in custom silicon and chiplet-based platforms, will manage the development flow for Mobilint, which contributes its neural processing technology and domain expertise in robotics.

Robotics places specific demands on accelerator design. On-device inference for robots requires high throughput at low power, deterministic latency for real-time control loops, and sufficient memory bandwidth within strict thermal and size budgets. The partnership pairs Mobilint's AI processor architecture work with SEMIFIVE's implementation and integration capabilities to meet those constraints.

For Mobilint, the contract gives the startup access to proven design infrastructure without building a full in-house chip engineering team. For SEMIFIVE, the deal extends its portfolio of AI semiconductor engagements and reinforces its position in Korea's edge-AI chip ecosystem, where government funding is steering investment toward on-device processing.

The robotics AI chip developed under the program will target deployment in robotic systems that need local inference for perception, navigation and control tasks. Further technical specifications, timeline milestones and volume targets have not been disclosed.

The agreement signals continued momentum in Korea's on-device AI semiconductor sector, where selected program participants are moving from concept phases into funded silicon development with established design partners.