Breaking the sound barrier: How AudioShake is teaching machines to hear like humans

Walk into a packed sports stadium, you hear the roar of the crowd, the play-by-play announcer, stadium music pumping through speakers, thousands of overlapping conversations. Or listen to your favorite song: the vocals, drums, guitar, and bass are woven together into a single sonic tapestry. Yet, without even thinking about it, the human brain picks apart these layered sounds, focusing on what matters while filtering out what doesn’t.

Machines? Not so much. AudioShake set out to solve this puzzle, and in doing so, they are fundamentally changing how we create, edit, and understand sound. The San Francisco-based AI company has built technology that can “unmix” any audio recording, separating it into individual, pristine components even decades after it was mixed together.

One of their most recent milestones was winning the top prize at the 2024 AWS re:Invent Unicorn Tank competition and becoming a key infrastructure partner with major artists like Green Day and companies like Disney Music Group as the next generation of multimodal AI.

Learn how AudioShake amplified its AI opportunity. Read the full case study

Over 330,000 have joined AWS Activate since its inception in 2013, accessing a range of resources to help build, grow, and scale their startup. This includes unlocking AWS Activate Credits, which can be used to help cover the costs of AWS services and solutions, from compute, database, and storage; to machine learning, containers, applications, and more. AWS Activate has already provided more than $7 billion in credits to startups, and yours could be next to benefit. Find out more about AWS Activate today.