Europa, OMR, Kinder-Karrieren – Live-Coaching mit Flo Heinemann

Heinemann explains why venture capital fits almost no startups

Value in this market keeps concentrating in a handful of AI and chip winners, and venture capital has to work around that. Florian Heinemann describes Project A's model, where deals come almost entirely from his own outreach because very few startups fit venture funding at all, and where private valuations rest on marks that hide losses until an exit. He discusses the IPO pipeline and argues private markets belong in an Altersvorsorgedepot.

Takeaways

• Global market value hits a record against GDP. Global market capitalisation rose by more than $1tn to a record $163tn, against estimated world GDP of $105tn to $110tn. Zschäpitz sets that against Warren Buffett's warning about stocks being worth more than GDP.
• The rally runs through chips alone. The S&P 500 gained 2.3% for a sixth straight week and the Nasdaq 100 5.5%, while the SOX chip index rose 11.1%. The DAX managed 0.2% and is down this year, with banks, healthcare, software and energy weak.
• Cloudflare's record fall may help model providers. Cloudflare fell 24% on Friday, its biggest one-day drop ever, on exploding token and inference costs. Zschäpitz says that pressure could be good news for Anthropic and other model providers.
• Project A has never invested from cold outreach. Heinemann says Project A has made about 160 to 162 investments and not one came from cold outreach, with 30 to 50% from its own approach. He talks to anyone who does not come across oddly.
• Venture capital suits almost no startups. Heinemann puts the share of US startups suited to venture funding at roughly 0.3%. He compares his job to a scout finding players who could reach Manchester City in five years and convince Pep Guardiola in five minutes.
• Private market marks hide losses until exit. Heinemann describes a stability illusion in private market marks and falling multiples. He is sceptical of semi-liquid retail private equity products but argues private markets belong in an Altersvorsorgedepot, ideally as a satellite spread across various funds.
• Signal shows concentrated tech is not inevitable. Meredith Whittaker, president of the Signal Foundation, told him language models fed on user data extend the advertising-based data economy. Signal, built by about 50 people, shows competition is possible. She cites a Nebraska conviction based on chat logs.

Heinemann's advice for the next generation

• Build on strengths rather than fixing weaknesses, which he says the German school system ignores by pushing everyone through one channel.
• Study MINT subjects or data science, and do not study law for its own sake.