What 1000 Kelvin is building
1000 Kelvin has evolved from its additive-manufacturing AMAIZE product toward a broader 'Physical AI Lab' focused on foundation models for laser processes. The company describes models trained around laser-material interaction, thermal diffusion, melt-pool formation, phase change and solidification, with the goal of predicting and controlling laser manufacturing behavior.
Where it sits in the engineering workflow
Metal additive manufacturing is highly coupled: geometry, scan strategy, laser power, material, thermal history and machine behavior all influence defects and distortion. Conventional parameter development relies on simulation, coupons and expert iteration. A physics-informed model can potentially predict problematic regions and adjust process parameters before a costly build is committed.
Why mechanical engineers should care
This is a compelling manufacturing-AI category because the output can influence physical process settings rather than only office productivity. Earlier collaboration with EOS integrated AMAIZE into an industrial additive workflow, and the company's current positioning extends the same physics-first idea to laser processes more broadly.
What it does not remove from engineering
Process models have to be qualified against the exact machine, material lot, optics and calibration state used in production. In regulated applications, any AI-generated parameter change may itself need validation and traceability. Users should distinguish between a model that predicts thermal behavior and a fully qualified production process window.
Questions to ask before adopting it
- Which laser processes and material systems are validated today?
- How is the model calibrated to a specific machine and optical system?
- Can every parameter adjustment be traced and reproduced for quality records?
- What physical coupons or in-situ monitoring are still required for qualification?
DIVE-LD engineering view
The most useful way to evaluate Physics AI for laser manufacturing / AMAIZE lineage is to put it inside a controlled engineering release process. Ask whether the output preserves design intent, whether assumptions are visible, whether another engineer can edit or audit the result, and what verification remains before manufacturing. AI can compress repetitive work and widen the design space, but responsibility for requirements, safety, standards, tolerances and final release still belongs to the engineering organization.
Sources and further reading
Public product information reviewed on 11 September 2026. Company statements about performance or capability are vendor claims unless independently verified.