HII Integrates Physical AI to Accelerate Shipbuilding

HII and GrayMatter Robotics signed a memorandum of understanding in early April 2026 to explore integrating GrayMatter’s Physical AI into HII’s shipbuilding operations. The MOU targets autonomous surface preparation, coating, and inspection — discrete, repeatable processes suitable for robotic automation — and was announced at a ceremony attended by Eric Chewning (HII EVP, maritime systems and corporate strategy) and Ariyan Kabir (CEO, GrayMatter Robotics).

“Physical AI” in this context means closed-loop, model-driven control architectures that combine computer vision, spatial reasoning, motion planning, and factory orchestration to operate robots in semi-structured manufacturing environments. GrayMatter positions its stack as a “Factory SuperIntelligence” intended to move beyond fixed, single-task industrial robots toward adaptable fleets that can perceive and act across varied assemblies. HII plans to fold those capabilities into its HYPR (High‑Yield Production Robotics) initiative to boost production throughput while preserving quality and safety.

• •Scope: The partnership will initially focus on autonomous surface prep, coating, and inspection, with potential expansion to broader unmanned and manned shipbuilding capabilities and scaling of unmanned-system production lines. The MOU also includes workforce training and integration workstreams. (HII press release / Yahoo)
• •Production targets: HII cited a 14% throughput increase in 2025 and stated an objective to pursue an additional ~15% throughput improvement in 2026, framing automation as a lever to reach those targets. (HII press release / Yahoo)
• •Leadership and positioning: Eric Chewning framed the collaboration as part of an “American shipbuilding renaissance,” while GrayMatter markets the underlying approach as enabling factory-level autonomy. The ceremony took place at GrayMatter’s Carson, California, HQ. (HII press release / Yahoo)
• •Practical fit: Industry reporting highlights that surface prep, coating, and inspection are more immediately automatable than craft-heavy tasks like weld qualification, which remain technically challenging due to variable geometry and skilled judgment requirements. (RedState, Manufacturing Dive)

This MOU is notable for practitioners working at the intersection of robotics, perception, and factory automation because it signals a large-scale, defense-sector use case for adaptive physical-AI control stacks rather than narrowly programmed manipulators. Naval shipbuilding combines large structures, variable geometry, and high-quality coatings/inspections — a demanding testbed for perception robustness, path planning in constrained environments, force/torque control for surface work, and systems integration with legacy tooling and production flows. Success here would validate approaches to fleet orchestration, simulation-for-deployment, and workforce upskilling that other heavy industries can reuse.

Expect integration challenges around sensor calibration in harsh environments, robust defect detection for coatings, cycle-time modeling, and safety certification for humans working alongside autonomous fleets. Workforce-training workstreams named in the MOU are pragmatic: automation at scale requires creating operator roles that supervise, maintain, and teach robots rather than simple job displacement. For ML/robotics teams, the program will demand reproducible evaluation metrics (throughput, rework rates, coating quality), digital-twin simulation to iterate control policies, and tight QA loops for naval standards.

• •Pilot outcomes and metrics: proof-of-concept throughput, quality, and reliability figures from early pilots. 2) Integration with HYPR: whether GrayMatter’s stack plugs into HII’s existing automation and orchestration layers. 3) Scope creep toward structural automation (welding/assembly) — success on surface tasks could lead to attempts at stricter craft automation. 4) Workforce programs and safety/regulatory frameworks for human-robot collaboration in shipyards.