Key results
The company
Yoona.ai
yoona.aiAI-powered design and automation platform for the fashion industry.
Result highlights
- 40-50% increase in sell-through rates
- Up to 87% design cost reduction for customers
- Time-to-market cut from 40 weeks to 9 weeks
- Time to insights cut from weeks to seconds
- Up to 93% CO2 reduction in prototyping
The story
A German fashion technology startup provides an AI platform that helps designers predict trends and reduce the textile industry's 1.2 billion-ton annual carbon footprint.
Rising training costs on a previous cloud provider made it difficult to scale the company's computationally intensive AI models. A lack of technical support prevented the team from configuring their infrastructure efficiently, limiting their ability to support complex GPU workloads.
The startup migrated to Google Cloud with partner Seibert Group, deploying Compute Engine and GKE to manage hundreds of Nvidia A100 GPUs. They optimized costs by reserving high-performance hardware for intensive training while using standard machines for inference. This architecture powers AI agents that analyze proprietary datasets to predict trends and generate design prototypes automatically.
Quotes
“The accelerator was the best we've ever done. The support and the mentors were amazing. It brought us closer to Google Cloud and showed us it was not just a technology provider, but a team we could work with for long-term success.”
Implementation partner
Seibert Group
seibert.groupHelped yoona.ai build a modern, flexible architecture to manage data and complex AI workloads.