Principles VenturePrinciples Venture

More powerful models, more powerful companies

Models will get 100x better. We back companies that get stronger because of it, where every leap in the model makes the product more powerful.

Principles Venture is a Brooklyn VC fund backed by a family office. We invest our own capital at Seed and Series A in US and Israeli AI infrastructure companies. Shlomo Landau founded it in 2025.

The thesis in one paragraph

Many AI products today are built around workflows that won't last. We invest in what stays: what AI can't run without, and what enterprises need to put it to work, whatever model or agent does the job. We back US- and Israel-based companies at Seed and Series A.

What won't last

When the App Store took off, there was an app for everything: flashlights, calendars, email. Then the phone built them in, and it didn't matter how many users those apps had.

AI is heading the same way. Anything that's a thin layer on top of a model, the model makers will ship as a button. That covers summarizing, writing help, meeting notes and chatting with your documents.

We look for the opposite: companies a better model makes stronger, not obsolete. They own data or rights others can't get, plug into systems that are hard to reach, build the infrastructure the models run on, or act in the physical world.

Our test is simple: does the product get stronger every time the model gets better?

The question we ask every deal

If the models get 10x better and agents do the work, is this still needed, or does it disappear with the workflow it serves?

Two made-up examples show how it cuts.

Take a tool that helps a person write sales emails faster. If an agent writes and sends the emails on its own, the tool has no one left to help. We pass.

Now take the system that stores the data the agent reads, or the check that tells a company when the agent's answer can't be trusted. Better models make those more needed. That's the kind of company we back.

What we mean by AI infrastructure

We use the term for two kinds of companies.

What AI can't run without

Every model depends on layers underneath it, no matter who built the model. The data has to be stored somewhere and served fast enough to use. In our portfolio, StorONE is the storage layer enterprise AI data runs on.

What enterprises need to put AI to work

A model on its own doesn't change how a company runs. The company has to know when to trust it. It also has to fit AI into the way its people already work. In our portfolio:

  • Neurath.ai tells enterprises when a model is guessing, whatever the model.
  • MatterMore maps how a company's people and its AI actually work together.
  • Galya is plug-in personalization any AI product can run on.
  • Xdysis puts AI to work at the core of the business, running workflows and automating processes.

Why only the US and Israel

We invest only in companies based in the US or Israel. A company based anywhere else is a pass, however good it is.

We have people in both places. Shlomo Landau runs the fund from Brooklyn, and Yona Kadosh, our venture scout, is in Tel Aviv.

How the thesis shapes how we invest

We invest at Seed and Series A.

We back few companies. When we do, we go all in.

We don't back what we don't understand, so we go deep before we commit. Once we've done the work, we decide fast.

For investors who come in with us: we source the deal, we do the work, and we put our own money in first.

Common questions

What is Principles Venture's investment thesis?
Principles Venture backs companies that stay needed as AI gets better: what AI can't run without, and what enterprises need to put it to work. It invests at Seed and Series A, only in US- and Israel-based companies. Its test for every deal is whether the company is still needed if models get 10x better and agents do the work.
What counts as AI infrastructure for Principles Venture?
Two things. The layers models depend on, like data storage. And what enterprises need to use AI in real work, like knowing when a model is guessing, seeing how people and AI work together, personalization, and running workflows.
Does Principles Venture invest in agentic AI infrastructure and tooling?
Yes, when it passes the deal test. Agents still need data to work from, a way to tell when they're wrong, and a way into a company's workflows. Those are the companies we back.
Which stages does Principles Venture invest at?
Seed and Series A.
Does Principles Venture invest outside the US and Israel?
No. It invests only in US- and Israel-based companies.

If you're building AI infrastructure in the US or Israel, send us your deck.