Substack passed 5 million paid subscriptions — what the number does and doesn't tell creators

Substack announced in March 2025 that it had surpassed 5 million paid subscriptions, a milestone reported across the industry press. It is a genuinely large number for a platform that launched its paid tier in 2017 — and it is also a platform metric, not a creator-income statistic. The 5 million counts subscriptions spread across many thousands of writers, stacked on free readership that runs far larger, and the company's 10 percent take on paid subscriptions shapes everything about how that revenue is distributed. Here is what the number actually tells a creator deciding where to build.

Licht Journal publishes information, not business or tax advice; figures are from company announcements and published terms.

What is inside the 5 million?

Subscriptions, first — a reader subscribing to two publications counts twice in some framings of the metric, which the company has acknowledged in its own postings. Second, growth concentrated at the top: the subscription economy on Substack follows a power law, with a small share of publications earning the majority of revenue — a pattern consistent with what published creator-economy research shows across every platform. Third, momentum from network effects: Substack's app, recommendations between writers and its notes feed all exist to move readers from one publication to another, which is why the company markets the network, not just the tool.

How does the money actually split?

Substack's published fee structure is simple: 10 percent of paid subscription revenue, plus payment-processing fees (Stripe's rates, roughly 3 percent). Writers keep the rest and own the content and, importantly, the mailing list — exportable at any time, per the platform's own terms. That list ownership is the single most consequential term for creators: it makes Substack exit-able in a way video platforms are not. Against that: no advertising revenue share, no algorithmic discovery comparable to YouTube, and discovery depends heavily on the writer bringing or earning an audience.

What can a creator realistically earn?

The honest answer is that the distribution is brutal and the median is low. The top individual writers on Substack have publicly reported seven-figure annual revenues — reported by the writers themselves, so self-reported and unaudited — while the long tail earns little or nothing. The workable math for a new publication: convert 2 to 5 percent of a free list to paid at a typical $5-8 monthly price, which means a free list of 10,000 readers at 3 percent conversion yields 300 paying subscribers, or $1,500-2,400 a month gross before fees. That is a side income, not a salary, and pretending otherwise sets creators up to quit early.

What generalizes from Substack's rise?

Three things transfer to any platform decision. One: reader-funded models reward consistency and voice, and penalize churn — publishing cadence is a business input, not a stylistic preference. Two: owning the export file is worth more than any discovery feature, because it converts platform risk into a migration cost. Three: network features (recommendations) measurably lift smaller publications — Substack's own data and writers' posts credit cross-recommendations as a top growth source, and the mechanism, one publication's readers becoming another's, is repeatable anywhere creators can trade audiences.

What does not generalize: the top earners' revenues, which describe outlier brand and audience accumulated over years — often imported from other platforms — and the assumption that any paid tier converts without a writer first publishing free work for months. Substack built the friendliest terms in the industry; it did not repeal the power law.