The SoftBank Vision Fund API Portfolio

About this paper

I read SoftBank’s Vision Fund portfolio the way a machine reads it — not the round size, the evidence. The Vision Fund is the largest pool of growth capital ever assembled for technology, a fund whose reputation is written in nine- and ten-figure checks into late-stage category leaders, and 87 of its companies ship a public API I could pull, parse, and score on the same rubric I apply to Stripe. This report is about those 87, because the interface is the part of a growth-stage company you can inspect independent of the valuation — and in the Vision Fund’s case, the interface tells a story its cap table does not.

The headline is not the one a megafund wants. SoftBank writes the biggest checks in venture and holds one of the thinnest ceilings on the board: of 87 API-bearing companies, exactly two clear the exemplar band — Polygon and Databricks — and Databricks clears it on a suspiciously shallow four-spec surface. The book averages a composite Kin Score of 50.0, dead center of the pack, with strengths that are cosmetic-forward — high discoverability and contract quality, big polished OpenAPI surfaces — and weaknesses that are structural: governance, operational transparency, security depth, and machine-consumability all lag, with idempotency at 9% and examples at 7% in a book heavy with payments companies. The real forward signal comes not from the marquee giants but from small fintech seedlings whose agent-readiness outruns their composites. This report reads that signal company by company.

What's inside

1. Executive summary — the biggest checks, the thinnest ceiling
2. The fund, and the book — the 87 API-bearing companies
3. Portfolio decomposition — wide, even, and uncentered
4. The composite Kin Score of the book — polished on the surface, thin under the hood
5. Agent-readiness of the portfolio — respectable, and misattributed to the seedlings
6. The peer-fund benchmark — the megafund in the middle, thinnest at the top
7. The crown jewels — provider-by-provider on Polygon, Databricks, and the strong band
8. Standards and the agentic turn — why the payments gaps stop being pedantry
9. Security, auth, scopes, and compliance posture across the book
10. The AI weight, the fintech tell, and the whitespace
11. The investable thesis — for SoftBank, and for everyone reading it

What you get for $750.00