This is the second article in the series about Ruby's future in the AI era, inspired by the keynote "Startups on Rails" by Irina Nazarova (CEO, Evil Martians) at RubyConf Thailand 2026. In Part 1 we covered the psychological barriers engineers face. Today — business and money.
There's a stereotype that building an AI product requires Python and nothing else. The numbers tell a different story: at SF Ruby Conference 2025, 25 startups demoed their products. 12 of them were AI-native companies, 5 were YC-backed, and 3 had already reached unicorn status. These aren't legacy companies hanging on — they're startups that chose Ruby in 2023, 2024, 2025.
Here are 4 standout examples.
1. Bolt.new: From Zero to $40M ARR in 8 Weeks
Bolt.new is a service for building and deploying full-stack web apps with AI. Founded in 2017, the company has raised $135M at a $700M valuation.
- Result: The team hit $40M ARR just 8 weeks after launch.
- The backstory: As CTO Albert Pai puts it: "It's an overnight success seven years in the making." The founders spent 7 years building a different product (StackBlitz). When the Bolt.new idea took off, the flexibility of Rails let them rebuild and scale to millions of users without rewriting the backend. "Rails was what caught my attention because it was just so easy to get started with. The ability to go to market fast has been a really key advantage."
┌────────────────────────────┬──────────────────────────────┐
│ Bolt.new Case ($40M ARR) │
├────────────────────────────┼──────────────────────────────┤
│ Stack │ Ruby on Rails + AI Agents │
├────────────────────────────┼──────────────────────────────┤
│ Time to $40M ARR │ 8 weeks │
├────────────────────────────┼──────────────────────────────┤
│ Funding │ $135M raised, $700M val. │
├────────────────────────────┼──────────────────────────────┤
│ Key factor │ Speed of hypothesis testing │
└────────────────────────────┴──────────────────────────────┘
2. Cora Computer: A Next-Generation Email Client
Cora Computer is an AI-native email client built by solo founder Kieran Klaassen. It launched in 2024 and was featured in The New Yorker. The product is built around AI-powered triage and auto-replies.
As Klaassen describes the experience: "I enabled it for my own account... The day after, it feels cleaner. We still had that urge to check and refresh, but there was nothing. It was just cleaner. It was quiet."
And: "I'm having a lot of fun. More fun than I've ever had building software."
3. Chat with Work: Enterprise Knowledge Search
Chat with Work, built by Carmine Paolino (creator of RubyLLM), lets you search and draft across all your work apps.
As Paolino puts it: "Chat with Work evolved far beyond just search. It's become more like Claude Code for your workplace. With Rails, RubyLLM, and the Async stack, we can build things that would require entire teams in other ecosystems." The setup: Rails 8 + RubyLLM, Gmail/Drive/Docs context, 54 seconds to a working AI chat.
4. Finn AI Inside Intercom: AI on 2.5 Million Lines of Code
Intercom ($1.275B valuation) is a major player in customer support. Hundreds of billboards for their new AI agent Finn AI are scattered across San Francisco, resolving over 1 million tickets per week.
The interesting part is under the hood:
- Finn AI was not extracted into a separate Python service. Python is used only for GPU workloads and model training.
- The product runs inside Intercom's existing Rails monolith — 2.5 million lines of Ruby code, 250+ engineers.
- Business logic, data, and integrations all stay in Rails.
As Eugene Kenny from Intercom put it at SF Ruby Conference: "2.5 million lines of Ruby code... 250+ engineers... In the age of LLMs, it's resilient." 14 years on Rails. Still growing. Still shipping.
And there are 8 more AI-native startups on Ruby: Simple AI (enterprise voice agents), Sixfold (AI insurance underwriting), Cactus (AI lead qualification), Superconductor (parallel coding agents), Bemi AI (AI customer support agents), Cleary (AI HR platform), LlamaPress (AI content platform on Rails), and CorePilot (AI-powered ERP for aviation).
Why Do Founders Choose Rails in 2026?
When it comes to an AI startup, the main thing is Time-to-Market. The winner isn't the one who writes perfect ML code (the model is accessed via API from OpenAI or Anthropic anyway), but the one who builds a usable UI, billing, auth, user logic, and database around that model fastest.
In 2026, Rails remains the best monolithic framework for rapid hypothesis testing.
➡️ Read Part 3: Why Ruby on Rails Is a Great Fit for the AI Era: Performance, Tokens, and Architecture