Cairn

Inspiration Up to 50% of US AI data centers face delays , costing operators $14.2M/month per facility. The bottleneck is invisible status. Highly skilled crews lose 5.5 hours every week just walking the floor or making radio calls to find out what work has already been done.

What it does Cairn delivers shared spatial awareness for data center commissioning via two capabilities:

On-Floor Headset Map: Automatically builds a live 3D map as crews move. Workers drop instant status markers visible to everyone else on the floor.

Remote 3D View: Off-site project managers see the live floor layout and can route workers dynamically without radio calls.

How we built it We built Cairn using the open-source SGA-AR framework (EasyVizAR). Because data halls are bright, climate-controlled, and static, we bypassed traditional AR limitations like low-visibility degradation. To handle network dependencies, we scoped a rapidly deployable, temporary onsite edge network.

Challenges we ran into Physical Friction: Standard AR hand-gestures fail on sites where workers wear safety gloves. We optimized our tagging workflow for voice and single-gesture inputs.

Site Security: Data center floors strictly ban cameras. We bypassed this by selling a vendor-to-contractor model where hardware is issued as approved, internal contractor equipment.

Network Gaps: Active builds lack stable connectivity. We had to bundle a temporary local site network into our COGS rather than relying on standard cloud access.

Accomplishments that we're proud of We designed an ultra-focused MVP with zero feature creep that directly solves coordination failure. We also built a highly pragmatic go-to-market model targeting electrical subcontractors to completely bypass slow general contractor procurement cycles.

What we learned We learned that the most critical question on a commissioning floor isn't "is it correct?" (which legacy AR solves), but rather "is it done, and who is on it next?" We also realized that construction accounting doesn't track rework codes , meaning our platform must actively instrument its own two-week baseline data during pilots to explicitly prove its ROI.

What's next for Cairn We are launching $75K paid single-hall pilots to explicitly measure and prove search-time reduction. Following that, we will build a native DCIM export integration to seamlessly hand off live spatial data to operations software at close. Long-term, we will scale into adjacent high-consequence verticals like semiconductor fabs and pharmaceutical cleanrooms.

Built With

• sga-ar