Skip to content
AI venture studioOfftilt is live on iOS & Androidchaoswing.co
ChaosWing
The studio

We start
companies,
not projects

A venture studio is a strange animal: part founder, part investor, part factory. Ours is at the founder end — every company here is one we started. Co-founding and funding other people's is where this goes next. Here's exactly how it works today.

Movement 01

Find the friction

We hunt for expensive, repetitive judgment calls inside real industries — the work people apologise for. Domain first, model second. If we can't explain the pain in one sentence, we drop it.

Movement 02

Build it ugly, fast

A working product in front of real users before anything is polished. Product, engineering and design run out of the studio itself, on an AI stack that has already been through production — so week one produces software, not a roadmap.

Movement 03

Run it, don't flip it

We keep operating what we start — support queues, audits, releases, the unglamorous middle. A venture is meant to outgrow us eventually, and when one does we'll bring in the operator who should own it. None has yet.

What would make us break the rule

Four things, in order

01
A domain you can't fake

Years inside the problem beat a clever wrapper every time.

02
Consequence

Somebody is losing money, time or sleep — measurably — right now.

03
A wedge, not a platform

One job done unarguably well, shippable in weeks.

04
Appetite for the boring part

Audits, brokers, regulators, support queues. That's where the moat is.

Straight answers

Do you invest, or do you build?

Today, only build — every company here is one we started and still run. Investing and co-founding with outside operators is the direction, not the current offer. Capital without hands was never the scarce thing anyway.

How early is too early?

There is no too early. A one-page note about a problem you've lived is a better start than a finished deck about a market you've read about.

Do you take outside pitches?

Not formally — everything in the portfolio, we started ourselves. But the door isn't locked. If you've lived a problem in one of our lanes, send it and you'll get a real opinion either way.

Who owns the company?

So far, we do — these are our own companies. When we do start one with an outside founder, they'll hold the majority and the title and we'll hold a studio stake earned by the build, agreed before anyone writes code.

Will you take on client work?

No. We used to look like an agency from the outside; we aren't one. Our upside comes from the companies we own, not from invoices.

What do you actually bring?

Product and engineering hands, design, and an AI stack that's already been through production in a regulated domain. Not a network, not a recruiting pipeline — we're not going to claim leverage we haven't built yet.

Bring us the problem you can't stop thinking about

Pitch the studio →Or see what we've started