Treadum
Every tire built to spec before it reaches the press.
Treadum runs the tire plant from compound to grade: Banbury mixing and compound control, tread and sidewall extrusion and ply calendering, green-tire building and splice control, cure-state control at the press, AI visual, X-ray and shearography defect inspection, and uniformity and balance optimisation — all tied together by an as-built tire twin and a genealogy record for every tire.
A tire is a dozen components built by hand-trained people and cured on a fixed recipe, and the plant learns about a mis-splice or a cord shift at the X-ray, long after the building drum that caused it. Nadeem and Rema are closing that loop at the drum and the press instead of reporting on it afterwards, and they built the audit trail first, which is what a safety-critical plant will actually let near its controls.
Treadum
- Seed · 2026Stage
- $1.5MRound
- Applied AISector
- London, United KingdomHeadquarters
- 8Team
Nadeem Ahmed
Co-founder & CEO
Nadeem runs the company and the commercial side: which cell goes first, what number it is judged on, and who at the plant signs off on the baseline. He started Treadum because tire plants already sense most of what goes wrong and still find out at final inspection, four hours and a few hundred tires too late.
Rema Sharma
CTO
Rema leads engineering: the factory-edge runtime, the vision, X-ray and shearography models that decide at the building drum, the as-built tire twin, and the bounded write-back into mixer, press and line controls. She is why every agent runs in shadow mode first and carries a genealogy record behind each action.
-
2026
We came in
$1.5M seed, with design partners across passenger and truck radial plants.
-
2026
Shadow mode
Splice, placement and cure-state baselines running against live building and curing data before any write authority.
-
2026
Building the team
Hiring edge-AI and controls engineers in London.
An illustrative 99SA Ventures profile — founder, team and figures stand in until the company shares its own.
Building AI that has to be right, not just impressive?
If your model gets graded against something real, we're the easy meeting to get.