TURING FOUNDRY

Most AI companies sell the factory. We keep the factory and own the output.

A software publisher operating a portfolio of niche apps, games, and websites — built and launched at a pace impossible before AI.

The shift

Software creation is becoming cheap. Software judgment is not.

For two decades, software companies were built around one product, one market, one roadmap. Writing code was expensive enough that focus was the only viable strategy.

AI changes the cost structure. When creation becomes cheap, the scarce layer moves upward — to knowing what to build, how to launch it, how to grow it, and when to keep or kill it.

Turing Foundry is built for that world.

The loop

A publishing loop, not a product roadmap.

  1. Mine demand

    Scan search intent, app reviews, communities, and user wishes to find underserved software needs.

  2. Generate products

    AI-augmented workflows turn validated demand into apps, games, websites, and growth experiments.

  3. Launch fast, kill faster

    Ship quickly. Measure retention, monetization, and acquisition cost. Most experiments die early.

  4. Own the winners

    Products that work become long-term software assets in the Turing Foundry portfolio.

The model

Voodoo-style publishing. Constellation-style ownership.

Voodoo finds and scales winners through rapid testing, creative iteration, user acquisition, and monetization — killing most experiments fast.

Constellation Software acquires and holds vertical software businesses for the long term, compounding cashflows across hundreds of niche assets.

Turing Foundry combines both. The difference: we don’t buy companies. We generate them.

Constellation acquires niche software. Turing Foundry synthesizes it.

Proof, not promise

What’s shipped, what’s repeatable, what’s the engine.

Shipped product

Anxinshe

A Chinese metaphysics and emotional-wellness app — exactly the kind of culturally specific market most Silicon Valley startups would never build for. That’s the point. AI makes long-tail, language-specific software economically viable for the first time.

200k registered users · 30k MAU · profitable
Repeatable branch

Camera-native games

A growing library of lightweight web games controlled by mouth, fingers, face, and body pose. Shared engine layer, new mechanic each launch — distribution through native social virality.

Demand engine

Wish-graph

Every wish submitted below becomes part of a growing dataset of unmet software demand. Not proof yet — the engine being built. Over time, the wish-graph itself may be Turing Foundry’s most valuable asset.

Cost structure and cultural reach let us operate niches Silicon Valley startups structurally can’t reach.

The arc

Where this goes.

2026

Ship 30 products across niche categories. Prove the loop.

2027

Ship 300. Operating cadence becomes the moat.

2028

A compounding portfolio of AI-native software assets, with the wish-graph as a proprietary map of unmet demand.

Wish for software

What do you wish existed?

Your wish becomes part of the demand graph. Strong signals — multiple people asking, clear pain, willingness to pay — go into the build queue.

mild annoyanceactively painful
Built by a Cambridge & Princeton alum, previously at Apple and Bridgewater Associates. Operating the portfolio hands-on.