Maritime AR Intelligence — for shipyards
The senior welder retires.
His knowledge stays.
Bulkhead is augmented-reality eyewear for shipyard trades — guidance, inspection, and recorded yard expertise, in their field of view while they work.
What Bulkhead does
What Bulkhead does
Welders, fitters, electricians, inspectors wear lightweight AR glasses. Bulkhead sees what they see — and helps.
Paperwork that closes itself by end of shift.
Every weld and inspection — verified, step by step.
Forty years of yard knowledge — hands-free.
The case
Why this product exists.
A shortage that clusters in the trades you cannot replace. New hires that quit before they are productive. Reject rates that drag the schedule. A starting wage thinner than fast food — four charts that frame why we built Bulkhead for the yard.
01 / 04
The shortage
The shortage
The Navy’s April 2024 45-day Shipbuilding Review put the floor at 174,000 new industrial-base workers over the next decade. The Department of Labor’s analysis of critical occupations — welding, soldering, frontline management — pushes the upper bound to 250,000 over the same period.
The shape of the shortage is consistent across both estimates. It clusters in second-tier journeyman ranks — the welder who has just enough years to run a panel but not yet enough to teach — and concentrates in trades whose certifications don’t transfer.
The American Welding Society projects 320,500 new welders needed by 2029, with 157,000 already approaching retirement. The training pipeline cannot close that gap in time.
Sources — Navy 45-day Shipbuilding Review (April 2024) · McKinsey / US Department of Labor (Jan 2026) · AWS Welding Workforce Data (2025) →Hover a cell to see trade × experience-tier detail.
02 / 04
The lost knowledge
Who is still on the yard
Each figure represents one in a hundred new hires this year. The Navy’s acting acquisition executive testified in March 2025 that 50 to 60 percent of new industrial-base workers quit inside their first year.
Hover a figure to meet them. Toggle the horizon to see who has stayed at year one, three, and five.
The yard pays training costs for all of them. GAO finds it takes three to five years to reach proficiency in the skilled trades — by the time a hire is productive, three out of four cohort-mates have left.
Sources — Brett Seidle (Breaking Defense, March 2025) · Eric Labs CBO testimony (March 2025) · GAO-25-108225 Navy Shipbuilding (2025) →On the yard at end of year one: 45 in 100
Source: Brett Seidle, Navy testimony — 50–60% first-year attrition
The yard has paid for training all 100
03 / 04
The verification gap
The verification gap
GAO’s 2025 report on Navy shipbuilding ties the workforce composition directly to the schedule: at one major shipbuilder, 57 percent of trade workers have fewer than five years of experience, and “quality problems and late discovery of rework” are explicit drivers of the delays. The lead Columbia-class submarine is twelve to sixteen months late in part for this reason.
On the floor, this shows up in reject rates. AWS Inspection guidance puts 1–2 percent at the productivity standard and treats 5 percent and above as “considered too high.” Reject rates rise as you walk down the experience ladder — and the experience ladder is bottom-heavy.
Procedure verification — drawn into the field of view, against the WPS — pulls the new-hire tier into the productivity band before the cost shows up on the schedule.
Sources — GAO-25-108225 Navy Shipbuilding (2025) · AWS Inspection Trends weld-rejection-rate guidance · NSRP 2017-442 (Newport News AR pilot) →04 / 04
The pay gap
The pay gap
The shipyard 5-year wage is competitive. The entry wage barely beats fast food. HII’s Kari Wilkinson testified in September 2025 that new hires earn $21 an hour at the yard, while regional fast food pays $18 — and the fast food shift is indoors and dry. The BLS national median for fast-food and counter workers is $14.92 an hour.
The yard wage can double inside 18 to 24 months for those who stay. Most do not. The Navy’s Brett Seidle told the House Armed Services Committee in March 2025 that 50 to 60 percent of new industrial-base workers quit inside their first year — before the divergence ever materialises. The yard pays for the training and watches it walk.
People can go do far less difficult things for just about the same money. Closing the entry-wage gap takes appropriations; closing the year-one survival gap takes a different tool.
Sources — Brett Seidle, Breaking Defense (March 2025) · Kari Wilkinson, Breaking Defense (Sep 2025) · BLS OEWS May 2024 (welder, pipefitter, assembler, fast-food) · Amazon corporate wage disclosure 2024 →Every one of those numbers describes a yard
operating right now.
The platform
Each capability, in detail.
Bulkhead runs on lightweight AR glasses. Trades wear them. The system sees what they see.
Paperwork that closes itself by end of shift.
Welders, fitters, electricians narrate as they work. Bulkhead captures, categorises, and files the record — weld parameters, hot-work permits, NCRs, daily reports — without them stopping what they are doing. By end of shift, the documentation is complete and the senior trade is not the one writing it.
Every weld and inspection — verified, step by step.
Before the bead is laid: torque specifications, weld procedure, gas pre-flow, joint geometry, lockout state, gas-free certificate — checked in the field of view, against the drawing. Steps cannot be skipped. The system makes deviation harder than compliance, and the rework cycle starts to shrink.
Forty years of yard knowledge — hands-free.
Senior trades record the way they actually do the work — torque sequences, fit-up corrections, the moment a bad weld becomes obvious. New hires retrieve the answer in their own words while both hands are on the equipment. The thirty-year welder’s knowledge stays on the yard after he leaves.
Who we are
Built from inside the yard.
Bulkhead was built by people who came up through marine engineering and shipbuilding, not through a case study. Our background spans yard operations, academic research in workforce learning, and production AI systems. We have read the rework reports, run the models on real production data, and spent time in the dry-dock where this work actually happens. We are not observers who found an interesting problem — we are engineers who hit the limits of what current tools could do and decided to build past them.
We stay small deliberately. Every decision is made by people who understand the apprentice’s first weld and the inspector’s last sign-off. That specificity is not a differentiator. It is the product.
Ready to put it in the
dry-dock.
We run early access pilots with shipyards who want to put this in front of their trades. Conversations are short and specific — no decks, no demos, just a direct discussion about your context.
We are currently talking to