Why We Invested in Cast AI

This is Part 3 of a 3-part blog series around the looming climate challenge: the massive surge in data center energy consumption driven by AI. See Part 1 and Part 2. Written by the team at G2 Venture Partners.

6 GW of Wasted Compute?

In the race to build data centers to support AI, we often overlook a fundamental truth: data center growth is, at its core, an energy problem. Computational demands have skyrocketed, placing strain on our data center infrastructure and energy grids. While much attention focuses on building new capacity, at G2 Venture Partners, we’ve recognized a critical opportunity hiding in plain sight: making more efficient use of what we’ve already built.

The U.S. currently has approximately 20 GW¹ of installed data center capacity. Yet, a significant portion of servers sit idle — 30% of all virtual servers and 25% of all physical servers have remained inactive for at least 6 months.² Further, active servers often run at far below their optimal utilization levels — ABB reports that servers typically operate at just 5–15% utilization.³

Why are servers so heavily underutilized? Consider the workflow of a DevOps engineer launching or managing an application. Their primary goal isn’t efficiency — it’s reliability. To ensure an application never goes down, engineers set resource requirements based on peak expected usage, with a healthy buffer, and then let the application run with this level of (mostly too high) resourcing. Periodic manual optimizations help, but DevOps teams tend to be stretched thin — at best, an engineer may check in to fine-tune the resource levels of the application every few weeks or months. The result? Massive resource waste, and this waste is an energy problem — a typical server consumes 30–40% of its maximum power even when idle⁴, creating a massive energy drain.

With Cast AI, DevOps teams can manage the resources of every application in real-time. Cast AI does what no human DevOps engineer could realistically do, right-sizing resources based on actual workload requirements every few seconds. On average, this reduces cloud costs by 20–30%, and in less optimized environments, savings can exceed 60%.

G2’s Investment in Cast AI

We are thrilled to announce our investment in Cast AI — a Kubernetes automation platform that enables customers to rightsize cloud computing resources in real-time, addressing cloud spend and resource utilization, thereby lowering the power consumed in data centers.

Cast AI’s platform automatically optimizes Kubernetes environments, ensuring that computing resources are matched to workload requirements. By dynamically adjusting resource allocation based on actual needs, rather than worst-case scenarios, Cast AI helps organizations eliminate waste. Cast AI’s technology not only delivers significant cost savings for customers, but also increases the reliability and productivity of applications. Crucially, it also minimizes the environmental footprint of data centers.

Our Investment Thesis

[1] Customers say Cast AI is the #1 cloud cost optimization vendor, delivering exceptional value.

Cast AI sets itself apart from other cloud cost optimization vendors through its technological sophistication, exemplified in the customer value it delivers. The team’s deep expertise has enabled them to develop optimization algorithms and automation features that outperform competing solutions in real-world deployments.

Cast AI delivers significant cost savings for customers. In conversations with dozens of customers, we found that even organizations that considered themselves well-optimized before implementing Cast still achieved ~20–30% reductions in cloud spend. For less optimized environments, savings routinely exceeded 60%.

But Cast AI’s value goes beyond simply cost reduction — its automation capabilities are particularly transformative for DevOps teams. By automating the manual implementation of its recommendations, Cast AI frees DevOps engineers to focus on more strategic, value-accretive tasks.

Accordingly, Cast AI plays an essential role in the modern cloud stack. Customers frequently highlight improvements in reliability and productivity through features like proactive scaling (which anticipates workload changes before they happen) and self-healing techniques that automatically detect and resolve issues in Kubernetes clusters, leading to higher availability.

[2] Cast AI is a trusted partner to enterprises across the spectrum.

Get G2 Venture Partners’s stories in your inbox

Join Medium for free to get updates from this writer.

Subscribe

Remember me for faster sign in

Companies of all sizes and industries rely on Cast AI, from technology leaders like Akamai to manufacturing giants like BMW to data specialists like NielsenIQ. Any company running Kubernetes with containerized workloads in private or public clouds can benefit from implementing Cast AI.

Notably, Cast AI is a trusted and reliable partner to enterprises. It has earned this trust through transparency, consistent performance, robust security measures, and exceptional customer support. Organizations that adopt Cast AI early in their cloud migration journey often expand its use, deploying it more broadly across their infrastructure over time.

[3] Cast AI is building the platform for DevOps automation.

In the last year, Cast AI has expanded beyond its core cloud cost optimization product. The company has launched new offerings in Kubernetes security, extended its reach into on-prem environments, and developed specialized optimization capabilities for AI workloads.

• Kubernetes Security: Automates vulnerability scanning and remediation, helping organizations maintain strong security postures without manual intervention.
• Cast Anywhere: Brings Cast’s optimization to all clouds and to on-prem environments, allowing companies with hybrid infrastructures to achieve consistent resource utilization across their entire computing apparatus.
• AI Optimizer: Optimizes LLM workloads by selecting the most cost-effective, performant, and secure AI model for a given task, reducing GPU costs and optimizing performance for enterprises running LLMs.

Each of these expansions reinforces Cast AI’s evolution to a comprehensive automation platform. By continuing to address adjacent challenges that customers face in their cloud stack, Cast AI further strengthens its value proposition — and there’s more to come!

[4] Cast AI is making a sizable dent in carbon emissions.

A consequence of reducing cloud costs for organizations is a reduction in provisioned compute resources. This has significant implications for energy consumption, and therefore carbon emissions. As Cast AI optimizes and rightsizes resources, it reduces the prevalence of idle and underutilized servers while also decreasing the need for new servers and data centers. When Cast AI helps a customer reduce their server count, those freed resources can be redeployed elsewhere — in effect, Cast AI increases the carrying capacity of existing data center infrastructure without requiring additional buildouts or provisioned energy.

Further, Cast AI’s climate impact extends beyond reduced energy consumption to embodied carbon. Each server produced generates ~2 metric tons of CO2e — by eliminating the need for additional servers, Cast AI reduces the production of new hardware and its associated emissions.

By 2034, we forecast that Cast AI’s platform will result in avoided carbon emissions of 8 megatons per year. To put that into perspective, this is equivalent to the annual emissions from about 1.8 million gasoline-powered cars or 1.6 million U.S. homes.

Automation is the Future

By investing in Cast AI, G2 Venture Partners supports a solution that addresses both the economic and environmental challenges of cloud computing and the data center industry. As workloads grow in size and complexity, the need for intelligent resource optimization will only become more critical. Cast AI is positioned at the intersection of this trend, helping organizations maximize the value of their cloud spend while simultaneously reducing their environmental footprint.

In a world where digital transformation and sustainability are increasingly intertwined, Cast AI represents exactly the kind of innovation we seek to support — technology that enhances business performance while driving measurable environmental impact, contributing to a more efficient and sustainable digital future.

If you or your organization are interested in Cast AI, book a demo here: https://cast.ai/#modal-book-a-demo.

¹ https://www.rystadenergy.com/insights/data-centers-reshape-us-power-sector

² https://www.computerworld.com/article/1680190/a-third-of-virtual-servers-are-zombies-2.html

³ https://new.abb.com/news/detail/66580/how-data-centers-can-minimize-their-energy-use

⁴ https://new.abb.com/news/detail/66580/how-data-centers-can-minimize-their-energy-use