Answers · Views and frameworks

What do investors check in an AI company's Series A?

FutureX Capital screens an AI application company on three things first: distribution, retention, and whether gross margin climbs as token prices fall. Then it asks one question: which of your features disappear when the next model ships?

The three screens test whether a company earns, and FutureX wants the evidence before the first call. Distribution: model companies hold the biggest entry points, and an app without its own channel is a feature in someone else's product. Retention: seat-subscription churn gets amplified by the next model, and only customers who stay for results count; FutureX benchmarks enterprise retention at 90%. Gross margin: inference cost for the same capability falls about 10x a year, and that gap should stay in the app's margin, provided prices keep falling; by September 2026 the frontier token price index had been flat for three months. For open-source projects it also checks stars and the path from community to paid.

Growth sets the backdrop. In a16z's September 2026 State of Markets, US venture-backed tech companies that recently raised show median revenue growth above about 60%, and new rounds still tend to require high growth; among top AI apps, the group starting at $1 million to $50 million ARR grew a median 423% year on year (report p.73 and p.76, approximate).

Moats test whether a company is hard to replace, and FutureX's diligence questions come before the financial model. First: which of your features disappear when the next model ships? A feature that exists because today's model cannot do something goes to zero with the next release. Second, for agent products: was the mechanism that brings a human back into the loop designed or skipped? Third, an upstream cut-off stress test: the share held by a single model supplier, the replacement cost, and the contract terms for switching models. The four moats FutureX accepts are taste, extreme efficiency in one workflow, combining several models, and process as product.

The speed benchmark: per company disclosure, Genspark reached $100 million ARR nine months after launch and $200 million in 11 months, a path that took five to seven years in the SaaS era; across the market, AI companies above $100 million ARR went from 3 in 2023 to 80-plus by April 2026 (Reported).

For founders →Report · a16z: Tech Drives About 76% of S&P 500 Profit Growth, Hardly a Bubble →Report · Agent Usage Grew Roughly 14x, but Pricing Power Goes to Data and Workflow Owners →

Related answers

Updated 2026-11-02. Facts and examples come from FutureX Capital's published research, public talks, and public reporting, graded per our site-wide standard (verified / reported / per company disclosure). Nothing here is investment advice or an offer.