Making music with machines
Some startups are born from spreadsheets. GenreX was born from sound.
Founded by Malcolm Yang and Yihao Chen, GenreX set out to make music generation not just automatic, but expressive, flexible, and tailored to the creative process. Their idea: let game developers, video editors, and creators generate high-quality, royalty-free music on demand, tuned to the vibe and timing of their project.
But turning that idea into a product meant more than just building a model. It meant moving fast, through the fog of research, into production, and all the way to a launchable app.
From research to production
GenreX knew their path would unfold in three phases: research, production, deployment.
In the research phase they wanted to explore and validate their AI architecture quickly, without blowing their budgets. As the models matured, they moved into production, training on larger and larger datasets which required scaling GPU needs as the models grew. Once training was over, they moved to deployment to serve those models reliably and cost-effective to real users.
The catch? Most platforms support one of those phases well, but not all three.
For GenreX, time was everything. Waiting on AWS quotas, managing environments manually, or switching platforms midstream would cost them not just time, but momentum.
“This platform is the perfect combination of tools for fast iteration and sanity-checking in the research phase—but can scale up for training a production model,” said co-founder Malcolm Yang.
From idea to audio in 4 weeks
Using Lightning AI Studios, the GenreX team got straight to work without setup delays, custom infra, or tooling churn.
They spun up GPU machines when needed, launched asynchronous jobs to test training variations, and moved fast between prototyping and production without rewriting anything. With BYOC (Bring Your Own Cloud), they used their own AWS account and budgets, while Lightning handled everything else: shutdowns, cost control, job orchestration, and file sharing via a universal file system.
“We validated our hypothesis in just four weeks for under $10K,” Yang said. “It would’ve taken us at least 70% longer without Lightning.”
Democratizing creative expression
Music is one of the most powerful forms of storytelling. Until now, it’s largely been reserved for those with the technical skills, expensive tools, or industry connections to create it.
GenreX is changing that.
By giving non-musicians, game developers, indie filmmakers, YouTubers, educators, and creators of all kinds, the ability to generate high-quality, royalty-free music on demand, they’re expanding who gets to tell stories with sound.
What once required a digital audio workstation, licensing deals, and a music degree now takes minutes and a creative idea. And as their tools integrate into platforms like Canva and Adobe Express, that accessibility is scaling fast.
With Lightning AI removing the infrastructure friction, GenreX didn’t just go to market 70% faster, they made it easier for the next million creators to follow.
Because when AI makes music creation frictionless, everyone gets to play.
Read more blogs and case studies:
• Decoding the brain by training 1000 AI models to predict movement
• Reducing carbon emissions faster at the MIT Sustainable Design Lab
• View all