Additive Industries and Materialise have released a new Next-Generation Build Processor for MetalFab metal powder-bed systems, available immediately. A company benchmark reduced preprocessing of a large, complex job from 46 minutes to six. That is an almost 87 percent reduction, but it is not a printing-speed record: it covers one measurement of machine-job preparation.

Where the build processor sits

A CAD model does not become laser motion directly. Build preparation orients and nests parts, creates supports, slices geometry, then generates scan vectors and machine-specific instructions. At the final, data-heavy stage, the build processor connects Materialise Magics preparation to MetalFab control.

Materialise's library lists the previous Additive Industries processor at version 2.6 for MetalFAB1. The new release is therefore not a universal vendor-neutral converter but a machine-family integration. The companies attribute its speed to multiprocessing and optimized algorithms. The geometry, computer hardware, repetitions and comparison protocol behind the published benchmark are not public.

MetalFab 420K functions move across the family

The release is more than a shorter wait. It is said to expose optimized multi-laser load balancing, beam diameter varied by vector type, ghost vectors intended to improve thermal stability and finer exposure-strategy controls. These capabilities were introduced with the MetalFab 420K platform. Variable beams and load balancing can in principle raise throughput, but the result depends on geometry, material, layer thickness and the qualified parameter set.

A ghost vector is an additional scan that may place heat in an area without a part contour to balance the thermal field within a layer. It is not a universal correction. Poor settings can alter cycle time and thermal history, so coupons and representative components must demonstrate the outcome.

Verified facts, vendor claims and open questions

The processor was announced on 21 October 2026, is immediately available to MetalFab users and integrates with Magics. Materialise's broader Next-Gen platform is available as a desktop application; cloud and on-premises server options are described as forthcoming. The 46-to-six-minute result, greater stability and cost-per-part gains remain vendor benchmark and vendor conclusions; no independent comparative test is public.

Open items include the exact release identifier, supported MetalFab configurations, licensing, reproducibility of existing jobs and which qualified parameter sets can use the new strategies without requalification. Faster file creation alone does not prove density, dimensional accuracy or surface quality.

Adoption checklist

First freeze the old software, processor, machine and parameter versions. Prepare the same small, large, multi-part and vector-intensive jobs in both versions. Measure load, slice and export times, file size and peak memory; compare layer count, vector count, laser allocation, estimated build time and warnings.

For physical regression, manufacture at least one geometry coupon, one surface coupon and one material-property specimen with the same powder and machine state. New laser allocation or ghost vectors require particular attention to multi-laser overlap zones and thermally sensitive features. Promote the version only after predefined acceptance limits pass, and test the rollback path as well.

Editorial conclusion

The significance is not the 46-to-six headline alone. Preprocessing can be a real bottleneck for large LPBF jobs, while machine-specific scan strategies influence both productivity and quality. This is therefore a meaningful software release. Its production benefit is proven only when it saves time on representative jobs while the job file remains traceable, physical output still conforms and qualification status stays controlled.

Sources and transparency

Release timing, benchmark and MetalFab capabilities come from TCT's 21 October report based on statements from both companies. Materialise's product page confirms the Next-Gen platform scope and availability; its library confirms the prior MetalFab processor. No independent performance test is available.