N+1 LabsDigital twins for factories

Fix your next build before you build it.

Factories find out what breaks after the tooling is cut. Every surprise costs weeks and another build.

Factory N+1 is a digital twin of your production line, built to be checked against real builds. It shows what will fail and the fix, before you build.

Book a demo See the prototype
An assembly line drawn twice: the real line on the left, its digital twin on the right as a glowing blueprint, with the press station flagged in red.
Real lineIts twin
01 · PREDICT

Run the next build in the twin

Thousands of virtual units, before tooling.

02 · FIX

See what fails, and why

The failure ranking and the change that fixes it.

03 · CALIBRATE

Correct it with every real build

Built to be corrected by each real build you share.

Prototype · simulated, uncalibrated · synthetic part

Change one dimension. Rerun 5,000 units.

A two-part enclosure: the cover snaps onto four plastic hooks. As it goes on, each hook bends. Bend a hook too far and the plastic is overstressed, so it can crack or lose its grip.

01 · The part
Exploded view of a plastic enclosure: a dark cover lifted above a grey base with four snap hooks on its walls. Snap hook
Cover and base, joined by four snap hooks. No screws.
02 · How it snaps
The cover slides down outside the base, its edge pushes the hook inward until the bend peaks, then the head snaps out into the window in the cover. inside outside 1 Cover slides on 2 Peak bend 0.1 mm gap 3 Snaps into windowThe cover slides down outside the base, its edge pushes the hook inward until the bend peaks, then the head snaps out into the window in the cover.1 Cover slides on inside outside 2 Peak bend 3 Snaps into window 0.1 mm gap
The cover’s wall comes down outside the base with a 0.1 mm gap. Its edge rides over the hook’s ramp and pushes the hook inward 0.8 mm (0.9 mm undercut minus the gap), then the head springs out into the window.
03 · The risk, at peak bend
At peak bend, the 8.0 mm hook reaches 2.74% strain at its root, over the 2.5% limit. The 8.5 mm hook, pushed the same amount, reaches 2.43%. 8.0 mm 8.5 mm 2.74% worst unit typical 2.25% 2.43% worst unit typical 1.99% 0%2%2.5% limit3%strain

Same push, two lengths.

A snap hook is a small cantilever. Bending is worst at its root, and root strain grows with the push and shrinks with the square of the length:

strain ≈ 1.5 · t · y / L2

The cover pushes both hooks by the same amount. Making the hook 6% longer cuts its root strain by about 11%, which brings the worst simulated unit back under the 2.5% limit.

Colour shows strain along each hook for the worst of 5,000 simulated units. Undercut, clearance and deflection drawn enlarged so they read.

04 · 5,000 simulated units
Peak hook strain for 5,000 simulated units. As drawn, 183 units exceed the 2.5% allowable strain. With the hook 0.5 mm longer, none do. As drawn hook length 8.0 mm 183 of 5,000 over the limit Hook 0.5 mm longer hook length 8.5 mm 0 of 5,000 over the limit 2.5% allowable 1.5%1.8%2.0%2.2%2.5%2.8% peak strain at the hook root, one bar = 0.05% Peak hook strain for 5,000 simulated units. As drawn, 183 units exceed the 2.5% allowable strain. With the hook 0.5 mm longer, none do. As drawn · 8.0 mm 183 of 5,000 over the limit Hook 0.5 mm longer · 8.5 mm 0 of 5,000 over the limit 2.5% limit 1.5%2.0%2.5%
Each unit gets its own part and station variation. Lengthen the hook by 0.5 mm and the whole spread moves under the limit.

Simulated and not yet calibrated against a real build. Synthetic part, not a customer’s. Strain is the model’s estimate for the worst of the four hooks. A full run takes about five seconds on one workstation. Explore five builds in 3D →

How a pilot works

Start with a build you already finished.

A blind backtest: we predict what went wrong without seeing your results, then you score us.

  1. 01 · SEND

    Share a past build

    CAD and tolerances from an EVT or DVT you finished in the last few months.

  2. 02 · PREDICT

    We predict, blind

    Our ranked failure modes and their causes, before we see your data.

  3. 03 · SCORE

    Compare with what happened

    Check our list against your real failures. About a week end to end.

Book a demo
Benjamin Shyong
Founder

Built by someone who has run the line.

Benjamin Shyong, founder and CEO. 17 years in manufacturing, from factory floors in Mexico and China to a decade of consumer hardware launches at Apple and Meta.

  • MetaLed the Ray-Ban Display AR glasses from prototype through mass production.
  • AppleRan iPad Pro operations from prototype to mass production, and stood up a new automation line for Special Projects.
  • Factory N+1Won NVIDIA’s Best Use of Nemotron, October 2026.

LinkedIn →benjaminshyong.com →

Where it goes

Station by station, then the whole factory.

  1. 01 · NOW

    Station twins

    Press-fit, snap-fit, screwdriving and pick-and-place.

  2. 02

    Line twins

    Whole lines, with contract manufacturers.

  3. 03

    Factory design

    Lay out lines and stations before they exist.

  4. 04

    Build-out

    Procure and stand up what the twin designed.

  5. 05 · CONCEPT

    MadeCell

    A transportable assembly cell, configured by the twin.

See it on one of your builds.

Book a demo Investors: request the deck

Or write to hello@nplus1labs.com