Zeit AI Raises €5 Million for the Mid-Market Data Layer

Zeit AI closed a €5 million seed round on 3 September 2026 to sell European mid-market firms an AI data engineer most of them cannot hire. Dealroom puts that cheque at about $5.8 million. The Munich company, founded in 2024 by Palantir alumni Leopold von Waldthausen and Marvin Bornstein, builds ZeitMind, an agent that connects business systems, cleans the tables, and turns a controller’s question into a live application.

The money is modest next to who wrote it. Hasso Plattner VC, the seed fund shaped by SAP’s co-founder, is in the round, as are former German finance minister Christian Lindner, Helsing chief technology officer Robert Fink, Meta board member Charlie Songhurst, 2014 World Cup winner Mario Götze and Martin Schoeller of the Schoeller Group, with Y Combinator, the University of Oxford’s Seed Fund, the Sequoia Capital Scout Fund and ACE Ventures.

The SAP Founder’s Fund Took a Stake

Hasso Plattner VC was set up in 2024 in Potsdam, beside the Hasso Plattner Institute, and talks about founders who strengthen Europe’s digital sovereignty. ZeitMind’s first named source system is SAP. The fund’s managing director, Dr Jens Schmidt-Sceery, did not pitch the round as a chatbot story.

Europe’s industrial base cannot outsource the decisions that run it. Zeit AI is one of the few companies making sovereign infrastructure the default rather than the compromise.

Jens Schmidt-Sceery, Partner, Hasso Plattner VC

Lindner’s name puts German industrial policy on the cap table. Fink, from Helsing, puts European defence-AI talent in the same room. Götze is the name that travels. Schoeller is old German industry. Charlie Songhurst brings a Meta board seat. Von Waldthausen has been plain about why they raised at all.

We could run this company profitably, but we are now supercharging the playbook we have put together over the past two years and expanding the team in Europe across Munich and an office in London that is to come.

Leopold von Waldthausen, CEO and co-founder, Zeit AI

He also said the firm is quasi-profitable and would not normally have run a fundraising campaign, and that the principal reason was to formalise ties with strategic advisers. The round buys a room as much as it buys runway.

WHO WROTE THE CHEQUE

• Hasso Plattner VC: SAP co-founder’s seed fund, Potsdam, sovereignty brief.
• Y Combinator: S24 batch, primary partner Nicolas Dessaigne.
• Oxford Seed Fund: university capital; von Waldthausen studied computer science there.
• Sequoia Capital Scout Fund: US scout money on a Munich data agent.
• ACE Ventures: European institutional slot beside the angels.
• Political and industrial angels: Lindner, Götze, Songhurst, Schoeller, Fink.

That mix is the product’s real distribution channel. Mid-market controlling teams do not browse Y Combinator demos. They listen to industrial families, former ministers, and the SAP orbit that already runs their ledgers.

What ZeitMind Does on a Controller’s Desk

Zeit AI’s own site sells a simple loop. A forward-deployed engineer puts the agent in front of finance, procurement or supply chain in weeks, and the company says no internal IT capacity is required. Ask in plain English, and ZeitMind connects SAP, ERP, CRM, HR systems and more than 600+ other sources, prepares the messy data, and ships a dashboard, report or application. Figures are meant to drill back to the individual transaction. Work is hosted in Germany, GDPR compliant, and processed in Europe only.

The company line on the old workaround is blunt: anything the ERP or the FP&A system does not support becomes an Excel stitch that costs ten days a month and never goes away. A controller describes the output; the agent is supposed to deliver the same afternoon.

WHAT THE AGENT IS SOLD TO DO

• Finance and controlling: Zeit AI cites 90% of reports automated and 370 hours a month given back, plus €450k from margin-risk alerts and €20 million of working capital pulled into one rolling view.
• Procurement and spend: The site claims €24 million saved in supplier talks, 1,500 parts consolidated across plants, and a 20% cut in quantified supplier risk.
• Supply chain: It lists €500k of overstock reduction, €200k saved on express freight, and €325k saved in dispatch planning.

Those euro figures are the company’s own case studies, not an audit. They are the specimen it puts in front of a CFO. The early Y Combinator demo is still the public walkthrough of the same idea, tables first, then a question in ordinary language.

Karl Allgaier, managing director of Allgaier Agrarhandel, said the team was able to use practical data applications on a daily basis after a few months. Claus von Heydebreck, president of Kolbenschmidt Pistons Brazil, called the tool effective at giving clarity in critical decisions. Maxime Bouchard, managing director of Jetfly, praised the user-oriented pace. The pitch is not a data lake. It is a finished object a plant manager can open on Monday.

Europe’s Data Teams Are 500,000 People Short

The sales story only works if hiring the human version is painful. IDC and the Lisbon Council, in a March 2025 facts-and-figures report for the European Commission, put the EU27 500,000-person gap in data skills for 2024, up 10.9% from 451,000 in 2023. The same study projected 535,000 in 2025 and 631,000 by 2030 on the baseline path.

THE EU DATA-PROFESSIONALS GAP

Bitkom, the German digital association, says 79,000 IT roles in Germany are unfilled now, down from 149,000 in 2023. That is a wider IT count, not a data-engineer census, and it has eased as the economy cooled. The people Zeit AI is replacing on the org chart are still scarce in Munich and still expensive relative to a mid-market headcount plan. Von Waldthausen’s line is that mid-sized firms need data engineering and cannot buy the talent at a price they will pay. ZeitMind is sold as that hire with no extra headcount.

They Left Palantir for the Spreadsheet Layer

Von Waldthausen spent 3.5 years at Palantir and ended up leading established customers in Germany, Austria and Switzerland. Bornstein spent five years there and owned the first deployments of what became AIP, across chemicals, railway infrastructure, media, utilities, insurance and medical equipment. Side by side for four years, they say they generated over €50 million in new project revenue and watched operational teams still move data by hand.

They left in summer 2024, joined Y Combinator’s S24 batch, and put the tool in front of controlling, procurement and supply-chain teams. The Y Combinator company page says they went from zero to €1M ARR in twelve months with six people, customers live in weeks, and no IT support needed. The same page now lists a team of 12 in Munich and London. Combined they claim over 15 years in data engineering and analytics.

FROM PALANTIR TO A TWELVE-PERSON SHOP

1. 2024: Founded in Munich after both men leave Palantir.
2. Summer 2024: Joins Y Combinator S24 and starts selling to real controlling teams.
3. First twelve months out of YC: Reaches €1 million ARR with six people.
4. 3 September 2026: Closes the €5 million seed and says a London office is to come.

The Palantir habit did not stay in California. Zeit AI is hiring a Forward-Deployed Engineer in Munich on a posted range of €70,000 to €230,000, plus backend and commercial roles. Von Waldthausen has written that half the team are former Palantir engineers, and that managers left Palantir, BCG and McKinsey to join. The autonomous data engineer still arrives with a human who has done this work at the high end. The new capital is earmarked to expand among mid-sized businesses and to hire the engineers who deploy ZeitMind.

Kolbenschmidt, Horsch and Jetfly Are Already Live

Named customers already on the record include the technology group Kolbenschmidt, agricultural machinery maker Horsch, agricultural trader Allgaier Agrarhandel, logistics firm Wahl & Co, Munich consultancy Stern Stewart & Co, and Luxembourg aviation company Jetfly Aviation. Von Waldthausen has said Zeit AI works with about 30 enterprise clients. Live work has been described in automotive, agriculture, aviation and logistics, the unglamorous systems where a wrong stock figure costs cash.

The install motion is a 24-hour proof on the customer’s own data, then a first six-month engagement that is supposed to prove savings before the account scales. The company tells buyers they will see a working application, not a slide deck, with every figure traceable to the transaction. That is a consulting sales cycle compressed into software, which is why the FDE job still exists.

It is also why the raise reads as a network round rather than a rescue. A shop that already has named industrial accounts, €1 million of ARR and a claim of quasi-profit does not need celebrity angels to make payroll. It needs doors into the next 300 Mittelstand controlling departments, and those doors sit with SAP-world capital, industrial families and a former finance minister.

The Consultants Who Bill for This Work

The quiet party in the round is the person who currently stitches SAP to a spreadsheet. Implementation partners, freelance data engineers and the in-house analyst who rebuilds the close every month are the labour ZeitMind is priced against. If the agent holds, that billable stitch work shrinks. If it slips on messy master data, those same people get called back in, which is the usual fate of “no IT required” tools the first time a plant code does not match.

Zeit AI does not escape that labour. It internalises it as forward-deployed staff and sells the output as software. Palantir already does a richer version of this for companies that will pay enterprise prices; the mid-market was the residue. SAP, for its part, has been folding its own AI assistants and data-cloud products into the same ERP estate. HPVC backing an overlay is a bet that the mid-market will buy a specialist agent faster than it will wait for the platform vendor’s bundle to reach a 400-person manufacturer.

German hosting is the political half of that bet. Lindner and Schmidt-Sceery are not in this round to watch European production data leave the continent. The product promise is that a controller in Bavaria can query the ledger in German, keep the rows in a German data centre, and skip the headcount fight with Munich salaries. The jobs board still wants the Palantir-shaped engineer who makes that demo survive contact with a live SAP client.

The London office is still a plan. The Munich FDE listing is already up.

Frequently Asked Questions

How Much Did Zeit AI Raise in Its Seed Round?

Zeit AI closed a €5 million seed on 3 September 2026. Dealroom places that amount in the 92nd percentile of a sample of 4,252 German seed deals and converts it at about $5.8 million. No lead investor was named. The company said it could have stayed profitable and used the round to lock in advisers and hire deployment engineers.

Who Founded Zeit AI and Where Does It Operate?

Leopold von Waldthausen (CEO) and Marvin Christopher Bornstein (CTO) founded the company in 2024. Von Waldthausen studied computer science at Oxford, published work on process mining and cryptography, quit a Yale MBA, and earlier sold weview to DemoUp Cliplister. Bornstein grew up around Berlin, spent five years in London, and has been writing software for 18 years. The firm is based in Munich, with a London office planned.

What Source Systems Does ZeitMind Connect To?

ZeitMind is sold as a connector to SAP, ERP, CRM, HRIS, databases and more than 600 other sources, including CSV files. Zeit AI offers a free proof of concept on a customer’s own data within 24 hours and says no internal IT team is required to start.

Which Companies Are Already Using Zeit AI?

Named accounts include Kolbenschmidt, Horsch, Allgaier Agrarhandel, Wahl & Co, Stern Stewart & Co and Jetfly Aviation. Von Waldthausen has said the firm works with about 30 enterprise clients, including some in the United States and two in Brazil.

Where Does Zeit AI Host Customer Data?

Customer data is processed in German data centres and in Europe only. The company says the service is GDPR compliant and that customer data is never used to train models, with results designed to be traceable to the individual transaction.