# How we helped Arc Boat Optimize for to be in a Generative AI World **The Problem** Arc Boats GenAI-driven engineering and simulation workloads required sustained, high-throughput compute across ECS and EC2. However, because the team relied entirely on on-demand instances and lacked guidance on how to leverage Savings Plans and Reserved Instances, their cloud spend was escalating unpredictably. This made it difficult to budget and allocate the resources needed to reliably run and complete their GenAI pipelines and expansion. Without external support, Arc Boat risked not only significant overspend but also the inability to fully execute and scale their GenAI workloads. They needed generative AI hands on technical support. **The Solution** Pump.co implemented a cost-optimized architecture to support Arc Boat’s Amazon Bedrock–powered GenAI engineering workflows by analyzing EC2 and ECS usage patterns, identifying stable compute baselines, and replacing volatile on-demand usage with strategically modeled Savings Plans and Reserved Instances. Pump.co integrated its automated commitment engine into Arc Boat’s AWS environment to continuously monitor compute demand across EC2, ECS, and Bedrock workloads, ensuring predictable costs and reliable capacity for GenAI simulations. Through targeted enablement sessions, Pump.co also trained Arc Boat’s engineering team on commitment management, cost visibility, and best practices for scaling Bedrock-based workloads—providing a stable and efficient foundation for completing their GenAI pipelines as they grow. **Success** Arc Boats saw achieved 62% reduction in costs across ECS and EC2 through compute savings plans. With a stabilized and predictable compute foundation in place, Arc Boat was able to fully execute and scale its Amazon Bedrock–powered GenAI engineering pipelines, removing previous bottlenecks caused by unpredictable cloud costs. The improved reliability and resource planning enabled Arc Boat to accelerate R\&D throughput and expand their GenAI workload capacity without architectural changes. As a result, the company met 75% of its OKRs for new engineering and simulation programs, driven by faster iteration cycles, greater infrastructure availability, and improved alignment between compute provisioning and GenAI model execution needs. Pump.co’s enablement sessions also strengthened Arc Boat’s internal FinOps capabilities, allowing the team to independently plan, forecast, and support new GenAI initiatives as the company grows. They began spending way more, month over month, and have exciting growth plans. **Analysis** Arc Boats spends 9,915.60, some of which include Gen AI workloads through G5s and G6s. This was a big win, since they joined with less than half of that spend, and have steady yet exciting growth plans in the future. GenAI workloads, including those built on Amazon Bedrock, require both scalable compute capacity and predictable cost structures to ensure uninterrupted model execution. --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://services.pump.co/generative-ai-services-and-stories/how-we-helped-arc-boat-optimize-for-to-be-in-a-generative-ai-world.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.