Case Study: Dashtoon cuts compute costs by over 60% with Simplismart

How Dashtoon cut their compute costs by over 60% with Simplismart

Dashtoon, a pioneer in online comic generation, faced significant challenges scaling its GenAI-powered Studio product. Their infrastructure, bottlenecked by inefficient autoscaling and high compute demands, required 15 GPUs to handle peak load and still resulted in high user latencies and unsustainable costs, prompting them to partner with Simplismart for a solution.

Simplismart deployed its universal backend and proprietary kernels to optimize Dashtoon's ML workflows. This transformation drastically improved performance and autoscaling speed, reducing the peak GPU requirement from 15 to 6 while achieving the target p50 latency of 30 seconds and cutting compute costs by over 60%.

Dashtoon