Yes or No Is Not a Diagnosis
Five different companies walk into an investment committee. A binary vote can carry exactly one of them. The other four get rounded off — to yes, or to no — and the rounding error is where the money goes to die.
The five are structurally different situations. One is ready for capital today. One is ready once two named conditions clear. One carries fixable execution breaks that capital would scale along with everything else. One is blocked by something outside anyone's control. One cannot clear its blockers as constituted. Each implies a different capital action — which is the case for verdicts instead of votes: Go·Deploy, Go·Milestone, Revisit·Execution, Revisit·External, Pass. Not gradations of enthusiasm. Five different objects.
The distinction that matters most to an IC is the one a vote destroys entirely: controllable or structural. A company facing controllable obstacles is a rescue with a defined path. A company facing structural obstacles is a different asset entirely, whatever the model says. Writing the same check into both is how the last cycle's portfolios got built.
A verdict needs an address
Every tension a read surfaces sits somewhere, and where it sits decides what can be done about it.
Some were built into the design — the plan wasn't buildable as drawn. A company that prices a product business while running a project business is not underperforming; it is executing a design faithfully, and the design is the defect. No amount of performance management reaches it.
Some are underdelivering now — the machine exists and produces less than the plan assumes. This is the layer every board pack already covers, and the only one most diligence can see.
And some sit under load — the break waiting at exactly the volume the capital is buying. Clean at eleven installs a week, breaking at thirty, because scheduling, logistics, and commissioning were never designed as one system. Current performance can be genuinely good while this layer is fully armed.
The under-load column is the one traditional diligence cannot see, and it is the one the check is about to create. The failure hasn't happened yet. The evidence has.
Why a single metric can't find it
The reason these breaks stay invisible is not that the data is missing. It is that each number, read alone, looks defensible.
So the read holds signals against each other: install rate against onboarding capacity. Engineering velocity against field quality. A growth plan against the cost base that has to produce it. What one executive believes about the platform against what another believes about the same platform. One signal against another, so the contradiction surfaces instead of averaging away.
That last word matters. A board pack is an averaging machine. Two executives who disagree profoundly about whether the company can scale produce, on the page, a single confident number — and the disagreement, which was the most valuable information in the company, is the thing that gets deleted in the formatting.
A bridge without a diagnosis is a pier
Which brings us to the most common improvised verdict in this sector: bridge financing written without a diagnosis.
A bridge, by definition, reaches a defined other side. Without the diagnosis there is no other side — there is only the hope that time was what the company was missing. And time was never what was missing. The serial version is worse: instalment after instalment, each one buying months, none of them buying the answer, while the position's strategic value decays toward the one buyer who was never in a hurry.
A pier, called a bridge.
The alternative is not more caution. It is a verdict with an address attached: what breaks, where it sits, whether the holder can move it — and, if the answer is yes, what has to be true before capital is safe.
The window has a clock on it
One structural situation deserves naming, because it is running now across storage, heat pumps, and charging: the commoditization cliff. A hardware lead built over a decade evaporates in a handful of quarters when a scale player floor-prices the category. The economics do not announce it; the cost curve simply crosses.
For a company in that position, the pivot toward software, services, and an ODM model is not optional — and it is fundable only inside a window, while the hardware margin still exists to pay for it. Nobody inside the company wants to propose that pivot, because it reads as proposing the end of what the company is. It has to be sanctioned from the cap table, and it has to be sanctioned while the money to fund it is still there.
A verdict that names this is worth more than a vote that approves the round. And a vote that approves the round without naming it has, in effect, funded the wall.
The full essay on Under Load goes further: all five verdicts with their capital actions, the serial-bridge arithmetic, the two localization questions, the Commoditization Cliff in full, and the withdrawal chain — what leaves Europe when the diagnosis never happens.