Florian Heinemann's Project A is raising a €325m V5 fund, up from €275m, as VC and PE fundraising hardens on weak returns.
Defence and European resilience make up 20 to 25 percent of Project A's portfolio, which backs only dual-use deals.
Project A's Trade Republic stake is 8.5 to 9 percent, with the broker rumoured at €12bn and profitable for nine quarters.
Heinemann on why he would not buy OpenAI at $500bn
Project A is a Berlin venture firm built on operational support for its portfolio, and it is scaling that into a larger pre-seed fund and an external consulting arm. Heinemann reads venture returns as hit-driven and scarce, so he stays early, holds Trade Republic and will not pay up for AI leaders. Europe's gaps in defence, energy and pensions, he argues, close only with private capital and industrial policy.
Takeaways
Project A raises €325m for pre-seed. V5 is up from €275m and keeps its pre-seed focus. Fundraising has become harder in venture and private equity because investors have not received the returns they expected. Over 80 percent of the base is institutional, including Nordic pension funds.
Project A opens its value-creation team. A11 has more than 120 people and now consults for external clients. Heinemann places it as more implementation-oriented than BCG, turning the operational support Project A built for its portfolio into a service other firms pay for.
Defence is a fifth of the portfolio. Project A only does dual-use deals, not armed systems, and the push was driven largely by partner Uwe Hoßmann, with Jack Wang in London. Heinemann expects European LP mandates to loosen and cites Palmer Luckey's Anduril.
Trade Republic is the standout investment. Project A holds about 8.5 to 9 percent of the broker, which is rumoured to be valued at €12bn and has been profitable for nine quarters. Quantum's €3bn valuation sits alongside it, and Heinemann calls venture returns hit-driven.
Heinemann would skip OpenAI at $500bn. He does not doubt OpenAI could become a trillion-dollar company, but says the risk-return is better elsewhere. He prefers AI buy-and-build on existing businesses to AI-first roll-ups, and Project A either invests early or misses out.
German pension reform needs private provision. Riester and Rürup are poor, expensive products, and Heinemann says the pension gap will not close without more equity-based private provision. He is sure Friedrich Merz knows this but finds it hard to push through.
What Heinemann expects next
Data-centre commitments will break someone. If the same computing power can be reached with far less, as DeepSeek suggested, a party carrying $50bn or $100bn of commitments will not be able to pay.
Europe needs its own Starlink and launch capability, and cheap electricity for AI and manufacturing. Even a face-saving peace in Ukraine would leave Europe investing more in resilience, from rare earths to AI sovereignty.
Agentic commerce and e-commerce should recover. Live entertainment and sport benefit from people wanting to meet, but Project A would not back the Icon League or Baller League because it wants a technological or data-based core.
The German pension gap cannot be closed without more private provision in equities. Savings banks would also benefit from offering such products, and Heinemann calls it a core Social Democratic concern.
Tap a name to go there · tap a line to see why they are linked