Portfolio — Applied AI

Powdrix

The print floor runs itself. You decide how far.

Seed 2026 · $1M Applied AI London, United Kingdom

Powdrix runs the additive floor from build file to finished part: build preparation and process-parameter control, in-situ melt-pool and layer monitoring, recoat and powder-bed control, thermal and distortion control, AI inspection of CT and vision data for porosity, cracks and lack of fusion, post-processing across heat treatment, HIP and support removal, robotic build-plate and powder handling, and an as-built part twin that checks density, geometry and takt before the build starts.

Additive is moving from prototypes to serial production, and the floors doing it are drowning in data and short of process engineers. One bad parameter set or one unstable melt pool can scrap a multi-day build, and the porosity that fails a qualification is usually found long after the part is made. Christopher and Jonathan are closing the loop at the machine instead of reporting on it afterwards, and they built the audit trail first, which is what a regulated floor will actually let near its parameters.

Powdrix

  • Seed · 2026Stage
  • $1MRound
  • Applied AISector
  • London, United KingdomHeadquarters
The founders

Christopher Hayes

CEO & Co-founder

Christopher runs the company and the commercial side: which additive floor goes first, what number it is judged on, and who signs off on the baseline. His ground is production additive manufacturing and what it costs when it goes wrong — the scrapped build, the failed qualification, the machine-days spent printing a part that was never going to pass — and he started Powdrix because floors still find porosity after the build is finished.

Jonathan Pierce

CTO & Co-founder

Jonathan leads engineering: the machine-edge runtime, the melt-pool and layer perception and control models, the as-built part twin, and the audit trail behind every agent. His expertise is edge inference and in-situ sensing, and he is why each agent earns its authority in shadow mode first, with the evidence for every action written down before it is allowed to touch a machine.

How it's gone
  • 2026

    We came in

    $1M seed to take the first additive floors from shadow mode to bounded control.

  • 2026

    Where it starts

    Scrapped builds and requalification first: baselines run against live builds before any agent is given write authority.

  • 2026

    Building the team

    Hiring machine-edge perception, controls and OT integration engineers, and people who have run an additive floor or a qualification programme.

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.