1000 Kelvin

1000 Kelvin is an additive-manufacturing-AI company headquartered in Berlin with US operations, focused on making metal 3D printing succeed on the first try. Metal powder-bed printing is notoriously unforgiving — a build can take hours and fail from thermal stresses that are hard to predict — and 1000 Kelvin uses AI to prevent those failures before they happen.

Its product, AMAIZE, is an AI co-pilot for metal additive manufacturing. Machine-learning models predict where a print is likely to fail and automatically generate machine-specific print files tuned to avoid those defects, rather than relying on trial-and-error test builds. AMAIZE has been integrated into the EOS additive-manufacturing suite and Autodesk Fusion, and a second-generation AMAIZE 2.0 entered early access in 2026. The AI is unambiguously the product — it is a defect-prediction and build-preparation model, not a marketing feature on a printer.

The honest framing: additive-manufacturing AI is a genuinely small niche, but within it 1000 Kelvin is a clear, load-bearing-AI example — the ML predicts and prevents build failures, which is exactly the expensive problem metal printing needs solved.