Application-layer AIhas no moat.

Every product built on a foundation model inherits that model's abilities, and so does every competitor. Features are prompts. Prompts leak. Interfaces converge.

When the model underneath gets better, it gets better for everyone at the same time. The advantage you shipped last month is the default next month.

They build data moats that are too thin to matter, raise rounds to buy time and ship the same capabilities in a different colour, with a pitch deck that explains why this wrapper is different from the last one.

Some go further and make their product worse on purpose. They train their own models that cost more and do less than what is available off the shelf, and they run business models that never pay for themselves. The product gets slower and the margins get thinner, just so there is something to call a moat.

They defend the product because they think the product is what they own. That works until the next model release.

Closer Ventures is a portfolio of companies and products that form one ecosystem around one customer: the sales organisation of the German Mittelstand. Field reps, inside sales, sales leaders. People who are overloaded, underserved and loyal to whoever helps them first.

Our companies and products share customers, distribution and infrastructure, and each one makes the others stronger. A company we start today inherits the relationships, the integrations and the lessons of every company before it. It does not start at zero. It starts where the last one got to.

Over the next five years we will build more companies and products, and buy where it brings traction and lets us expand the ecosystem even faster. Many SaaS startups from before this wave are running out of time, while their customers still need what they were promised. We buy them for those customers and bring them onto what we build. It is the fastest way to grow a portfolio around one customer.

A competitor can copy what we shipped last month. They cannot copy what we ship next week. Every release puts us closer to the customer, and every customer conversation sharpens the next release.

That loop is what we protect: the speed at which a small team turns a customer's problem into software that solves it.

Every product built on a foundation model ends up with the same features, the same interfaces and the same demo. A customer comparing two of them is comparing two versions of one thing. A better product stops deciding who wins.

What does not converge is everything around the product: the customer relationships, the integrations, the distribution and the trust a customer builds with the companies behind it. A single product can ship fast. It cannot ship a whole ecosystem.

Our companies and products are built to work as one. A customer who starts with one gets the next on top of a relationship and an integration that already exist. Every company we add makes the rest more valuable, and every customer we win is a customer for the next product.