Skip to content
ProofofIncome
Subscribe
Menu
← All cases

Nick Walton (Latitude)

AI game / app · last checked 15 Oct 2026

AI Dungeon

● UnverifiableARR⏳ Figure from May 2025 · 17 mo old

Old figure: from May 2025 (17 months old). The business may look very different today; we show it with its date rather than as a current number.

TL;DR

  • Figure: $7M+ ARR (ARR · May 2025 · third-party leaderboard, reshared by the CEO)

    ≈ $583,333/month in USD — $7M+ ARR ÷ 12.

  • Proof level: Unverifiable. The only figure comes from a third-party leaderboard (reshared by the CEO, but not published by the company) and there is no primary filing or payment data. Our methodology rates that Unverifiable.
  • Our verdict (1/10, opinion): Not an indie case: a company dating from the 2019 era with years of iteration and a team under 20. The figure is secondary.

Reported figure

$7M+ ARR

Proof level

Unverifiable

Agentic-ness (opinion)

2/5

Replicability (opinion)

1/10

The story

Nick Walton leads Latitude, maker of AI Dungeon, a long-running AI game; the company predates 2024.

  1. May 2025 — "$7M+ ARR, 8M users, profitable, fewer than 20 people" — Henry Shi's Lean AI leaderboard (posted 22 May 2025), reshared and endorsed by CEO Nick Walton on LinkedIn

⚠ Later third-party estimates disagree; no primary filing found.

Business model

  • What it sells: subscriptions to an AI text-adventure game.
  • To whom: consumers/gamers.
  • Pricing: Research in progress (subscribe page returned 404).
  • Acquisition: Research in progress.

AI stack

ToolRoleMonthly cost
ClaudeInternal AI agents (per the shared post)Unknown

Game models: Research in progress.

Proof

Sources:

Why "Unverifiable": the only figure comes from a third-party leaderboard (reshared by the CEO, but not published by the company) and there is no primary filing or payment data. Our methodology rates that Unverifiable.

The Proof Report

Like this case? One new business each week

Weekly: one new AI business with proof-rated revenue, plus updates on the ones we track.

Form not loading? Subscribe on Beehiiv →

Can't find the confirmation email? Check your spam or promotions folder.

By subscribing you agree to our Privacy policy. Unsubscribe anytime.

Agent blueprint

Our suggestion — not how the founder did it

  1. Step 1

    Story agent

    Runs in-game narration and characters

  2. Step 2

    Moderation agent

    Filters unsafe content

  3. Step 3

    Ops agent

    Monitors model costs per user

Human in charge

A person approves money, public posts and key decisions. We don't estimate how much human work this saves.

Agentic-ness today (our estimate): 2/5 · Difficulty: Very hard

Risks & limits

  • Model costs.
  • Content moderation.
  • Not a beginner case.

Our verdict

Editorial opinion

1/10

Replicability today: 1/10. Not an indie case: a company dating from the 2019 era with years of iteration and a team under 20. The figure is secondary. (Editorial opinion.)

Similar cases

Figures as reported by the cited sources on the dates shown. Verdict, ratings and blueprint are editorial opinion. Not financial advice. No income guarantees.