We built the Business–Investment Boundary Atlas to make the assumptions behind that question visible. It compares four linked views of the same scenario, with selectable axes and equations you can inspect and change.
A robot can identify an apparently profitable deal. That leaves a harder question: can this particular operator build a sustainable business around repeatedly doing it?
We built the Business–Investment Boundary Atlas to make the assumptions behind that question visible. It compares four linked views of the same scenario, with selectable axes and equations you can inspect and change.
A transaction that is too expensive for a person to research and execute can become practical when a robot lowers the operating cost. But the difference between human and robot cost is only the beginning. Fees, losses, owner labor, software, compliance, capital lock-up and fixed overhead still matter.
There is no universal percentage edge that makes an activity a business. A 0.001% edge earns $10 on $1 million of turnover before overhead and the cost of capital. The same percentage can be meaningful at a very different scale, or disappear within estimation error.
The atlas compares the same transactions under human and robot costs, calculates a break-even margin, and discounts the economics over an explicit horizon. Expected inflation enters the nominal capital hurdle once; it is not subtracted again.
A good historical backtest does not establish that profitable opportunities will always exist. A gradient-boosted model and a foundation model face the same economic test: do the selected, executable deals produce a net edge on unseen data, including losing selections?
The simulator distinguishes the observed average from an evidence-adjusted margin. You can vary effective independent sample size, outcome volatility, an additional uncertainty haircut and the rate at which the gross edge decays. These are simplified assumptions, not a substitute for measured validation.
“Always available” becomes a question about a window of time. How likely is at least one qualifying opportunity in the next month? How long could a drought last? How many deals can the operator actually execute while capital is tied up?
The same deployment can be manageable for one operator and fragile for another. Below the minimum ticket, a deal may be inaccessible. With too little reserve, an otherwise positive strategy may stop during a loss sequence or before the first settlement.
The atlas separates available starting capital from capital allocated to deals. It includes reserve targets, an initial settlement delay, a deal-drought stress and a funding-interruption approximation. Unrelated wealth stays outside the venture's capital-charge denominator. The reserve policy is explicit, so you can test the cost of committing more cash as well as the resilience it buys.
The economic solver changes deployment size while holding starting wealth, measured edge and opportunity flow fixed. It cannot solve an underfunded scenario by assuming that more money or better predictions appear.
An operating business can earn less than its cost of capital for years. An investment strategy can have a repeatable edge. Neither fact alone determines a legal classification.
That is why the atlas displays operating character, economic viability, market capacity and regulatory readiness separately. Its perceived and precedent-informed lenses are transparent heuristic models. The precedent weights have not been fitted to a labeled case dataset, and the economic inputs are illustrative rather than measured P34 results.
Use it to ask which assumptions change a conclusion, where a strategy becomes capital-constrained, and what evidence would be needed to justify moving a boundary.
Explore the atlas and its full methodology →
The interactive page contains the formulas, research references and limitations. It is a research tool, not a forecast of returns or a determination of legal status. The source code, the model specification and the 13 numerical checks are public at hyperc-ai/business-investment-atlas on GitHub.
Version 2026-10-22. A research note: nothing here is a forecast of returns in any market.
Research notes as we write them, open problems as we hit them, model and agent-tool releases and member findings. Free, and you can leave whenever you like.
By subscribing you agree to the Privacy Notice. Unsubscribe anytime.
The working paper, benchmark methodology, notebook results and the limitations we state rather than hide.
See the evidence →Research, open problems, releases and what we are learning about computable markets.
All notes →Research note — 22 October 2026. The Business–Investment Boundary Atlas is a transparent theory explorer with illustrative assumptions; its outputs are not measured P34 results.