Growth Score Index, 2026 Q2: Only One Company in 897 Clears the Bar

What this index measures

Every research report in the Zen Horizon library answers the same ten growth-investing questions, adapted from the discipline long associated with Baillie Gifford: moat durability, unit economics, market ceiling, founder alignment, the next growth engine, and ultimately the path to a five-fold return in ten years. Each question is scored 0–10, and the sum is the company's Growth Score (0–100). Only companies with all ten questions scored make the index; where we cover several listings or share classes of one company, the latest report on the canonical listing counts once. The scoring standards and process are documented on our methodology page, and the full live ranking is on the Growth Board.

This first quarterly edition covers 897 companies, with data pulled from the production database on 2026-07-03. About 90% of the underlying reports originated in Chinese and the rest in English; the library skews toward companies we considered plausible growth candidates in the first place, so every figure below describes our coverage universe rather than the market as a whole.

The headline: quality is common, hypergrowth is rare

The median Growth Score is 43/100. The middle 80% of companies sit between 36 and 50, the lowest score is 21, and the highest is 66. Exactly one company in 897 clears 65: NVIDIA, at 66 (report), followed by Palantir at 61 (report), Sungrow Power at 60, and Atlassian and CATL at 59.

A distribution this compressed at the low end is deliberate. The framework asks whether a company can plausibly return five times its current value in ten years, which compounds to roughly 17.5% a year. Most good businesses fail that test even while passing every quality check, and the scores say so plainly.

Where the points die

Averaging each of the ten questions across all 897 companies shows exactly which hurdles separate a good business from a scoreable growth story. The three highest-scoring and three lowest-scoring questions:

The pattern across the full universe: questions about what the company already is (moat, customer dependence, margins) average above 5, while questions about what it would still have to become (a doubling of revenue, an unpriced edge, a credible 5x path) average 2.4–3.8. In our coverage, the scarce asset is a believable forward story that the market has also failed to notice — scarcer than quality itself.

Industry medians: software leads, regulated yield trails

Among the 24 industries where we cover at least 10 companies, the top and bottom of the table:

The 13.5-point spread between software and packaged food is the structural gap between businesses that can grow without proportional capital and businesses that grow with GDP. Utilities and banks in our library often score respectably on moat and customer dependence, then lose the growth questions almost entirely.