INXM has closed a €5.7 million Pre-Seed round and stepped out of stealth mode. The Berlin startup builds an AI Process Execution Engine that turns plain instructions into enterprise work that runs the same way every time. That focus on reliable execution is what sets it apart in a crowded enterprise AI market.
Cherry Ventures led the round, with Redstone, Angel Invest, and Linden Capital joining in. The round, announced on 3 June 2026, funds INXM’s first enterprise deployments and deeper work on its core product.
The Problem and the Bet:
Companies have poured billions into AI over the past few years. Most of that spending bought dashboards, copilots, and pilot projects. Very little of it produced workflows that run in production without breaking.
INXM aims straight at that gap. The pitch is simple and refreshingly concrete. Stop asking a model to interpret every step at runtime. Use AI to design the process once, then run it like tested code. That shift is the whole bet.
Compiled AI is the core idea behind the company. You describe what you want in natural language. A large language model turns that request into deterministic, enterprise-ready code. Then you run the code to get your outcome.
This approach keeps the flexibility of plain language and adds the reliability of tested software. The output stays predictable on every run. Compliance teams get a full audit trail. Operations teams get the same result, in the same format, each time they execute a process.
Meet the Founders:
The team comes from hard engineering backgrounds, not just software demos. Alex Oelling, the CEO, served as Chief Digital Officer at both Isar Aerospace and Volocopter. He built the digital infrastructure behind some of Europe’s most demanding aerospace programs.
Matthias Kainer, the CTO, worked beside Oelling at both companies on mission control and launch orchestration. Jesper Bylund and Kamil Klüber complete the founding group, bringing experience from n8n and Thoughtworks. People who help ship rockets and air taxis learn exactly what breaks under real conditions, and why.
The Product and Who it’s for:
Orchestrator is the flagship product. It plugs into the tools a company already runs, such as ERP, PLM, MES, QMS, email, and approval systems. It coordinates work across those systems rather than replacing any of them.
INXM calls a saved workflow a Plan. Once a Plan delivers good results, teams store it and rerun it for the same output every time. Picture month-end reconciliations, vendor onboarding, and reporting handled from start to finish. One early example is a 200-page RFP analyzed, risk assessed, and drafted into a bid in minutes.
INXM targets enterprise and Mittelstand operations, the mid-sized firms that form Germany’s industrial core. These businesses carry heavy process complexity and run lean engineering teams. They want results without multi-year implementation timelines.
The company says deployments deliver measurable outcomes within months and need no extra engineering staff. Everything runs in Europe with full data ownership and local deployment. It is GDPR compliant and developed in Germany, which carries real weight for enterprise AI buyers bound by strict governance rules.
What Comes Next:
Investors framed INXM as a new category, not another tool. Cherry Ventures founding partner Filip Dames said the team reframed the problem by making AI executable instead of simply smarter. He called Compiled AI a new architectural paradigm.
Redstone’s Michael Brehm pointed to the founders’ track record, noting that building rocket engines and air taxis teaches a rigor most software teams never face. Angel Invest’s Jens Lapinski added that enterprises do not need more disconnected AI tools; they need AI that runs governed, repeatable processes across the systems they already own.
This funding sets INXM up to ship its first production deployments and prove Compiled AI on real factory floors. Compiled AI as a term is new, and INXM is positioning itself to define it. For founders and operators tracking enterprise AI, the round signals where the category is heading, from clever demos toward governed, repeatable execution.
For anyone tracking where enterprise AI is actually going, INXM is a name to file now and revisit in 12 months.
