Cristóbal Valenzuela

Runway

Runway: Von NYU-Forschungsprojekt zu Hollywoods KI-Video-Tool

$90M ARR

Current revenue · as of June 2025

Sacra

4 years

$100K ARR

Communities

Growth Channel

Co-Gründer

Founder

Cristóbal Valenzuela

Co-Gründer•Technical•USA

Commitment

Full-time

Experience

First-time

Product

Runway

KI-gestützte kreative Tools für Videobearbeitung, Generierung und visuelle Effekte, die von Hollywood-Studios und Kreatoren verwendet werden.

Type

SaaS

Industry

AI / ML

Model

Freemium

Marketing Strategy

How Cristóbal acquired customers

Growth Channel

Communities

Also Used

Twitter / XMundpropaganda

Tech Stack

Tools used to build Runway

Custom AI modelsPythonTensorFlowWebGL

The Full Story

TL;DR

Cristóbal Valenzuela startete Runway als NYU-Forschungsprojekt und baute es zur KI-Plattform hinter Oscar-preisgekrönten Filmen wie "Everything Everywhere All at Once" aus.

Ich habe mit Runway angefangen, während ich an der NYU studierte. Es begann als Forschungsprojekt – ein Algorithmus zum Stilisieren und Kolorieren von Bildern in Photoshop.

Der Forschungs-First-Ansatz

Bei Runway waren wir immer forschungsorientiert. Wir wenden nicht einfach existierende KI-Modelle an – wir führen unsere eigene Forschung durch und entwickeln neue Techniken. So bleiben wir der Konkurrenz voraus.

Produktmarktkompatibilität finden

Der echte Durchbruch kam, als wir erkannten, dass professionelle Kreatoren – Filmemacher, Videobearbeiter, visuelle Künstler – dringend zugängliche KI-Tools brauchten. Hollywood-Studios geben Millionen für VFX aus, die wir helfen konnten zu rationalisieren.

Die Oscar-Verbindung

2023 gewann "Everything Everywhere All at Once" Best Picture bei den Oscars. Der Film verwendete unsere Rotoskopie-Tools. Dieser Moment zeigte der Welt, was KI-gestützte Filmproduktion erreichen konnte.

Die Zukunft von Video bauen

Jetzt bauen wir Gen-2 und darüber hinaus – KI, die ganze Videoclips aus Textanweisungen generieren kann. Wir bearbeiten Video nicht mehr nur, wir erschaffen es.

Gegründet: 2018 (an der NYU)
Eingeworben: $200M+
Hollywood-Partnerschaften: Disney, A24 und mehr
Team: 100+ Forscher und Ingenieure

Key Takeaways

Forschungsgestützte Unternehmen können dauerhafte Wettbewerbsvorteile in KI aufbauen

Universitätsprojekte können zu Milliarden-Dollar-Unternehmen werden

Die Ausrichtung auf professionelle Kreatoren schafft Glaubwürdigkeit für den breiteren Markt

KI sollte kreative Arbeit verbessern und unmögliche Dinge möglich machen

Originally published onCristóbal Valenzuela (Personal Website)

Founder proof brief

A one-page proof brief turns Cristóbal's Runway playbook into your next test.

See what it covers before you sign up: the pattern worth copying, the channel to test, and the risk to avoid when you adapt this story.

Pattern

$100K ARR

Channel

Communities

Output

Action checklist

What a proof brief covers

A one-page brief that turns Cristóbal's pattern into a testable plan
Channel bet: start from Communities
A risk check before copying the Runway playbook

See what a proof brief covers

See what it covers first. Create an account only if you choose to unlock.

Inspired by Cristóbal's journey? Generate a business idea in the AI / ML space using AI and real founder data.

Milestone Journey

Cristóbal achieved 3 milestones on the path to $100K ARR

Erster Kunde

6 months

December 2018

Avg: 1 month

+1 year to next milestone

$10K MRR

$50,000

2 years

June 2020

Avg: 1 year

+2 years to next milestone

$100K ARR

$600,000

4 years

June 2022

Avg: 2 years

4 years

Total journey time

Milestones achieved

Against every documented journey

Cristóbal Valenzuela reached $10K MRR in 2.0 years, faster than 20% of the 295 documented $10K journeys (median 12 months). The amber bar is this journey.

See the full distribution

Cristóbal's Path to $100K ARRPremium

The journey, decisions, and context behind this milestone

How Cristóbal actually did it

7/7 funnel stages sourced

Reconstructed from Cristóbal's own published sources. Quoted claims are the founder's words, verified verbatim; anything marked as our analysis is exactly that.

Where the idea came from

Side project spinout

Runway began as Cristobal Valenzuela's ITP graduate thesis project at NYU, exploring how to make machine learning models simple enough for artists to use without deep ML expertise.

“My ITP thesis project, Runway, is an invitation to explore these ideas.”

Machine Learning En Plein Air: Building accessible tools for artists (personal Medium)

Validation

Built first, validated after

MvpValidated

Launched a free public beta of the Runway tool, announced to his NYU ITP network and Hacker News, rather than pre-validating with landing pages or surveys.

The May 2018 Medium post confirms the tool was live and open to the public in beta, free of charge.

“Runway will be available for free and it's in beta now.”

Machine Learning En Plein Air: Building accessible tools for artists (personal Medium)

First test users

Online communityFree testers@ITP_NYU (Twitter)Hacker News (Show HN)

Cristobal recruited the first beta testers directly by tweeting the project to his NYU ITP network and asking interested people to DM him.

“Looking for beta testers, DM”

Cristobal Valenzuela launch tweet (@c_valenzuelab)

First paying customer

UnknownFree to paid upgrade

Runway charged from the beta onward for cloud GPU time, not for the app: signing up was free and came with $10 of GPU credit, after which running models remotely cost 5 cents per minute, so the earliest paying users converted out of the free beta pool once that credit ran out. No source names the first paying customer.

Per Runway beta pricing page, August 2019 snapshot (archive.org)

Acquisition channels, and whether they worked

Twitter / X · @ITP_NYU networkPre-launchWorked

A single tweet tagging his NYU ITP program, showing a demo video and asking people to DM him to become beta testers.

“Excited to share the project I've been working on in recent months”

Cristobal Valenzuela launch tweet (@c_valenzuelab)

Word of mouthScalingWorked

Hundreds of thousands of users gained since the first product release, credited to organic, product-driven growth rather than paid acquisition or a sales team.

Valenzuela attributes Runway's user growth to product quality driving organic, bottom-up adoption rather than deliberate sales or marketing motions.

Per Cristobal Valenzuela, CEO of Runway, on rethinking the primitives of video (Sacra research Q&A)

What didn't work

Product scopeLaunchCost: time

Runway's first product was a general-purpose directory of hundreds of machine learning models, an app store for ML, and by the end of 2020 the founder was describing that general-purpose phase as something Runway had moved past.

“Today, Runway is so much more than a general-purpose platform for AI”

Building Impossible Things (Runway company blog, byline Cristobal Valenzuela)

Usage on the model directory concentrated on one thing, video, with editors and filmmakers asking for workflow automation rather than a broad catalogue of models to browse.

Per Cristobal Valenzuela, CEO of Runway, on rethinking the primitives of video (Sacra research Q&A)

What they did instead: Rebuilt the product around machine learning for video editing, starting with rotoscoping and expanding into inpainting, depth estimation, transcription and audio tools.

Tech choiceLaunchCost: time

Runway shipped as a downloadable desktop app where models ran through Docker on the user's own machine, including an experimental bring-your-own-GPU mode that required Linux, a CUDA-capable NVIDIA card and nvidia-docker.

“The release of Runway 0.8.0 allows you to run machine learning models with your own GPU hardware.”

Announcing local GPU support (Runway company blog, byline Cristobal Valenzuela)

That setup put the product behind a local hardware and Docker install most of the artists and designers it was built for did not have, and within a year the site was pointing visitors at a browser version instead of the download.

Per runwayml.com homepage, August 2020 snapshot (archive.org)

What they did instead: Moved model execution to Runway's own cloud GPUs and rebuilt the product as a web app.

How it changed as they grew

At launch: 3 people (Three NYU ITP graduate students: Cristobal Valenzuela, Alejandro Matamala, and Anastasis Germanidis.)

Today: 21 people (21 people full-time, roughly tripling headcount from 7 the prior year, alongside tripling revenue year over year.)

Series a plus · $8,500,000 · 2020-12

What broke as they grew

Compute and infrastructure cost for running ML models at scale is an ongoing constraint the company manages deliberately, choosing not to pass those costs directly onto users.

Per Cristobal Valenzuela, CEO of Runway, on rethinking the primitives of video (Sacra research Q&A)

Sources (6 in the founder's own words)

NYU Tisch Honoree Speech (personal site) · founder-authored · 2026-04
Machine Learning En Plein Air: Building accessible tools for artists (personal Medium) · founder-authored · 2018-05
Runway raises $8.5M Series A (Runway company blog, byline Cristobal Valenzuela) · founder-authored · 2020-12
Cristobal Valenzuela, CEO of Runway, on rethinking the primitives of video (Sacra research Q&A) · third party
Cristobal Valenzuela launch tweet (@c_valenzuelab) · founder-authored · 2018-04
Building Impossible Things (Runway company blog, byline Cristobal Valenzuela) · founder-authored · 2020-12
Announcing local GPU support (Runway company blog, byline Cristobal Valenzuela) · founder-authored · 2019-08
Runway beta pricing page, August 2019 snapshot (archive.org) · third party · 2019-08
runwayml.com homepage, August 2020 snapshot (archive.org) · third party · 2020-08

Unlock Cristóbal's Full Journey

See the complete breakdown: launch strategy, validation methods, startup costs, expert analysis, replication playbook, and more actionable insights.

Upgrade to Premium

Instant access to all founder journeys

Frequently asked questions

How much does Runway make?

Runway reports $90M ARR as of June 2025. Runway $90M annualized June 2025 (up from $70M end-2024). $308M Series D April 2025 at $3B+ valuation. Source: Sacra.

What is Runway?

KI-gestützte kreative Tools für Videobearbeitung, Generierung und visuelle Effekte, die von Hollywood-Studios und Kreatoren verwendet werden. It operates in AI / ML.

How long did it take Runway to reach $100k arr?

Cristóbal Valenzuela took 4 years to reach $100k arr with Runway.

Was Cristóbal Valenzuela a solo founder?

No. Cristóbal Valenzuela built Runway with co-founders.

What marketing channel did Runway use to grow?

Runway grew primarily through Communities.

What industry is Runway in?

Runway operates in AI / ML.

Share this story:

Explore similar stories

$100K ARR Communities AI / ML Co-Gründer

Enjoyed this story?

Get more founder journeys like this delivered to your inbox every week.

Join founders learning from real success stories

Founder stories

Browse Stories

The Full Story

TL;DR

Cristóbal Valenzuela startete Runway als NYU-Forschungsprojekt und baute es zur KI-Plattform hinter Oscar-preisgekrönten Filmen wie "Everything Everywhere All at Once" aus.

Ich habe mit Runway angefangen, während ich an der NYU studierte. Es begann als Forschungsprojekt – ein Algorithmus zum Stilisieren und Kolorieren von Bildern in Photoshop.