The promise of metal additive manufacturing has always been undercut by the cost of failure. A single flawed part, discovered only after a 40-hour build and a trip to the CT scanner, can erase the economic case for the entire process. NexusAM, a London-based startup, is betting that the answer isn't a better scanner, but a smarter layer of software that watches the printer work.
Founded in 2024, NexusAM is developing an AI-driven, in-situ monitoring platform for laser powder bed fusion (LPBF) and similar metal 3D printing processes. The company's core proposition is machine-agnostic: it integrates with existing industrial printers, using sensor data to detect defects as they form and predict final part quality in real time. The goal is to replace or drastically reduce the need for slower, offline non-destructive testing like X-ray CT [PERPLEXITY SONAR PRO BRIEF].
The software wedge into a hardware problem
Most in-situ monitoring solutions are tied to specific printer manufacturers or require proprietary hardware sensors. NexusAM's wedge is a pure software play. By positioning itself as a drop-in SaaS layer, it aims to sell into existing fleets of industrial printers without requiring new capital equipment. This targets a key bottleneck for service bureaus and manufacturers scaling metal AM: the qualification and inspection process that currently acts as a throttle on throughput.
In a 2025 podcast, CTO Sebastian Larsen claimed the technology can reliably detect defects down to 200 microns during the build [TCT Magazine, February 2025]. That resolution, if proven at scale, would put it in the range needed to catch the porosity and lack-of-fusion flaws that plague high-value components in aerospace, medical, and energy applications. The full commercial launch is planned for 2027, suggesting a long runway of development and customer validation ahead of a general release.
The competitive field and the Creator Fund bet
The market for additive manufacturing software is crowded, but the in-situ quality assurance segment is still defining itself. NexusAM enters a field with several established and emerging players, each with a different technical approach.
| Competitor | Primary Approach | Notes |
|---|---|---|
| Phase3D | Optical coherence tomography (hardware+software) | Provides 3D volumetric data of the melt pool [TCT Magazine]. |
| Addiguru | Multi-sensor monitoring & AI analytics | Focuses on process stability and anomaly detection. |
| Additive Assurance | Acoustic monitoring | Uses sound signatures to identify defects [TCT Magazine]. |
| Jentek Sensors | Electromagnetic sensor arrays | Legacy provider of in-process inspection for composites and AM. |
| Applied Optimization | Multi-physics simulation & monitoring | Leans on predictive modeling alongside sensor data. |
NexusAM's differentiation rests on its AI stack and its software-only integration model. The company is backed by The Creator Fund, a UK early-stage VC focused on student founders, though the size of the investment is undisclosed [The Creator Fund]. The solo-founder structure with Larsen as CTO points to a technically focused early team, with the commercial and scaling challenges likely still ahead.
The long road to 2027
The three-year timeline to a full launch is both a risk and a signal. In a fast-moving software space, it provides ample time for competitors to solidify partnerships or for printer OEMs to develop their own integrated solutions. However, it also reflects the deep technical and regulatory hurdles in industrial quality assurance. Proving statistical equivalence to traditional NDT methods for flight-critical or implantable parts is a years-long endeavor of data collection and certification, not just algorithm development.
The technical breakdown is straightforward in concept, complex in execution. The system must fuse data streams from optical, thermal, and possibly acoustic sensors mounted on or near the printer. An AI model then correlates subtle anomalies in this data,a localized temperature spike, an irregular melt pool signature,with the subsurface defects a CT scan would later reveal. The real challenge is building a training dataset vast and varied enough to generalize across different machines, materials, and geometries.
At scale, the failure modes are predictable. Sensor calibration drift across a factory floor could introduce false positives. A novel alloy or geometry outside the training distribution might produce a silent failure. And the ultimate customer adoption hurdle may not be detection accuracy, but integration complexity; convincing a plant manager to wire a third-party software suite into a multi-million-dollar production cell is a different sale than a dashboard license.
NexusAM's bet is that the economics of reduced scrap and faster throughput will outweigh these integration fears. If its AI can reliably catch a 200-micron flaw before the build plate cools, it won't just be selling monitoring software. It will be selling confidence, and in metal additive manufacturing, that is the most valuable material of all.
Sources
- [TCT Magazine, February 2025] #239 NexusAM CTO Sebastian Larsen on enabling 'better functionality & economics' with in-situ monitoring | https://www.tctmagazine.com/podcasts/239-nexusam-cto-sebastian-larsen-on-enabling-better-functionality-economics-with-in-situ-monitoring/
- [The Creator Fund] NexusAM - The Creator Fund | https://thecreatorfund.com/case-studies/nexusam/
- [TCT Magazine] Phase3D, Addiguru & Additive Assurance among five finalists in ASTRO America in-situ monitoring challenge | https://www.tctmagazine.com/additive-manufacturing-3d-printing-news/software-and-simulation-news/phase3d-addiguru-additive-assurance-among-five-finalists-in-/
- [Crunchbase] NexusAM - Crunchbase Company Profile & Funding | https://www.crunchbase.com/organization/nexusam