KREA: Scaling AI-Powered Media with io.net

KREA is a bleeding-edge AI image and video generation platform that incorporates 2,500+ machine learning models to give users incredible creative control over creating and editing media. But KREA’s functionality requires serious GPU bandwidth to process extraordinarily complex computations.

With growing demand from clients like Nike, Apple, and Meta, @krea_ai recently hit a wall trying to scale their operations. The global shortage of GPUs has hindered many, including KREA, as the surge of interest and demand for AI products grows along with its expanding user base.

KREA got the compute power it needed to meet its ongoing growth, and a reliable, cost-effective way to:

• Access a large pool of high-performance GPUs.
• Scale its operations rapidly to meet growing demand.
• Maintain cost-efficiency for sustainable growth.

KREA partnered with io.net to take advantage of our decentralized network of over 300,000 GPUs, providing:

• Access to NVIDIA A100–80Gs GPUs: High-performance GPUs optimized for AI workloads, deployed on Kubernetes for seamless integration with KREA’s existing infrastructure.
• Cluster Scalability: A vast network of GPU clusters able to dynamically scale KREA’s demand for computational power to meet any surge in user activity.
• Actually Affordable Compute: Our GPUs are ~60% cheaper than the market average, letting KREA minimize operational costs and focus on their product.

The current global GPU shortage is slowing down many AI projects during a key moment for growth and expansion. By partnering with io.net, KREA gets access to global GPU resources today. This strategic choice has allowed KREA to focus on developing its product and expanding user base. For AI projects like KREA, access to scalable and cost-effective decentralized compute is table stakes to win. Partnering with io.net provides this access in a way that is easy and reliable.