How Railway says no to 90% of ideas and keeps winning | Jake Cooper (Founder, CEO)

Summary · 5 min read

Jake Cooper founded Railway after leaving Uber in 2020, initially building a box of databases for his side projects. That project became Railway, a cloud platform for deploying applications, databases and agents; the description says it serves millions of developers, has around 40 people and is nearing $100 million in annual revenue. Cooper remains Railway’s solo founder and CEO.

Jake Cooper says Railway grew by repeatedly addressing its biggest constraint, from the economics of its free tier to product, positioning and operating systems. He describes a lean, remote company that rejects most proposals, hires autonomous people and uses a six-week onboarding process to build trust, clarity and judgment. Cooper argues that AI will expand software production while increasing demand for compute, permissions and cloud-based agent workflows, and says Railway is moving more development into the cloud. He also links his leadership growth to letting go, giving precise feedback and protecting time to think.

Minute by minute

1. 2:09

   Railway rebuilt free-tier economics. Cooper says Railway started in 2020 as database hosting and later expanded up the stack. The description says the company was losing about $500K a month in its free-tier era; Cooper says Railway shut off free access while it built bare-metal capacity for two or three years, then restored it when the economics worked.

2. 6:04

   Growth came from repeated iteration. Cooper describes a year-long trough after Railway changed its free tier, followed by compounding growth from late 2022 and a further lift as models and cloud agents improved. Rather than point to one breakthrough, he says teams should tackle the biggest company problem each week and move it forward; Railway is adding agent experiences to its dashboard, CLI and MCP.

3. 9:41

   Test your own answer first. Cooper asks his team to reason through a problem before studying how Stripe, Cursor or a former employer solved it, because copying can constrain experimentation. He says Railway checks product decisions with users and internal dogfooding, while looking to other companies for inspiration in areas such as marketing.

4. 14:30

   Solo founding trades consensus for speed. Cooper says he did not set out to be a solo founder: after leaving Uber, he built a database box for side projects and posted about it on Twitter while looking for a job. He received no job offers but a few angel checks, then committed to building the company. He argues that two co-founders can get stuck in disagreements, while solo founding brings speed and leaves him as the final person responsible for unresolved problems.

5. 19:58

   Focus and systems create leverage. Cooper says Railway had about 40 people and was just shy of $100 million in ARR, and that he wanted to build with a small group of highly motivated people. The team rejects roughly 90% of ideas, asks why each proposal needs to happen now and plans quarterly. Cooper says teams should first do work manually to understand it, then build systems and automate repeatable tasks.

6. 25:17

   Player-coaches build operating leverage. Cooper says Railway looks for gritty, autonomous leaders who can do the work, build systems and coach others, rather than expand teams by default. After an early hire revealed information a background check would have surfaced, the company began running background checks. He says it looks for people who can move the business forward and change its trajectory.

7. 28:50

   AI raises the talent bar. Cooper says remote hiring lets Railway recruit globally and argues the company should not be stingy with compensation: he evaluates it about every half-year and may change it each quarter. He says AI raises the ceiling on what engineers can produce, so candidates can bring prepared work or build during an interview; Railway reviews their choices and patterns, not just the output.

8. 35:29

   Onboarding moves from trust to judgment. Cooper says the old grab-a-shovel approach stopped working as Railway’s codebase and context grew. Its six-week onboarding starts with two weeks of straightforward tasks to build trust, moves to problems with multiple possible solutions to build clarity, and ends by asking what the business needs. Quarterly planning still determines which work the team prioritizes.

9. 42:22

   Remote work needs deliberate systems. Cooper says remote work lets Railway hire globally and pass work across time zones, but requires systems for sharing context; the company added meeting transcription bots that attach context to planning items. Railway’s values spell HARD: honor, acceleration, reliability and directness. Cooper says honor is hardest, including when it means resisting overselling or publicly explaining incidents.

10. 48:40

    Stillness creates room for strategy. Cooper says his phone goes on do not disturb from Friday evening until Sunday afternoon, and he uses the time to journal and recover without returning to a screen. He argues that working 996 brings at most 20% more cycles, and usually less, while deliberate downtime can help founders find high-impact changes.

11. 54:46

    AI shifts workloads into the cloud. Cooper estimates that 100 times more people writing 100 times more software could mean 10,000 times more software, putting pressure on inference and compute; he also flags permissions and controlled access for agents. He says agents can automate the path from development through production, monitoring and support. He expects more development work to move off laptops and into cloud environments, which he sees as an opportunity for Railway.

12. 1:01:28

    Founders must let go to scale. Cooper says he has had to let go of engineering and other responsibilities so others can own them, and that learning to manage managers is a separate challenge. He says managers need to adapt their feedback to the person and situation. He warns that founders can bottleneck a company by holding on to work they enjoy or already do well.

13. 1:09:20

    Burnout signals an input-output mismatch. Cooper frames burnout as a mismatch between effort and results, and says he checks basics such as sleep, diet, exercise and caffeine when he feels tired. When an initiative stalls, he recommends isolating one variable to determine whether the issue is effort, method or execution. He says time away helps him spot mistakes and return with a clearer plan.

14. 1:12:22

    Purpose matters more than startup advice. Cooper says most startup advice is overrated because each company’s circumstances differ; he recommends taking in advice but developing your own judgment. He urges founders to decide whether a company’s purpose is worth committing to for a long time, arguing that sustained work can compound more than repeatedly switching paths.

Published Oct 13, 2026.