Cloud costs spiral as decentralised access fuels uncontrolled resource use: CloudKeeper CEO

The shift to the cloud is reshaping how businesses operate, but managing cloud costs remains a complex challenge. CloudKeeper, founded in 2017, provides cloud cost optimisation solutions to enterprises and startups, helps organisations control spending through a combination of rate negotiation, usage monitoring, and architecture improvements.  

In a conversation with TechCircle, CEO Deepak Mittal, breaks down the key issues enterprises face in cloud cost management, the growing impact of Generative Artificial Intelligence (GenAI), and how CloudKeeper’s tools and recent acquisition are addressing these challenges. He also offers a clear view of where the cloud industry is headed and what to expect in the coming years. Edited Excerpts:

What are the key challenges that enterprises and startups face in managing their cloud spending effectively?

When companies begin their cloud journey, the initial focus is usually on migration, security, reliability, and uptime. These are top priorities. Cost, however, is often overlooked.

First, many organisations don’t anticipate how quickly cloud costs can spiral. Second, their traditional IT procurement models, largely CapEx-based and centralised, don’t apply in the cloud, which operates on a flexible, on-demand model.

In the cloud, everything is available instantly. Hyperscalers offer hundreds or thousands of services, accessible to anyone in the organisation. That means anyone can deploy high-cost resources like top-tier GPUs or load balancers with a few clicks, often without oversight or approval.

This creates major governance issues. There's little control over what’s being used or whether it’s necessary. Organisations also struggle to understand the cost structure of cloud services and lack the tools to monitor usage in real time, hourly, daily, or monthly.

Ultimately, this is a multi-faceted challenge. The existing mindset doesn’t fit, a shift from CapEx to OpEx and from centralised control to decentralised access. There’s a lack of cost awareness, inadequate governance, and missing tooling. It’s a people issue, a cost issue, and a governance issue all at once.

How is your company using GenAI to drive innovation and improve offerings?

With the rise of GenAI, the industry is seeing both new opportunities and challenges. For most enterprises and startups, the availability of advanced GenAI tools from cloud providers like AWS, Google Cloud Platform, and Azure is a positive development. These tools offer powerful capabilities, but they also come with increased costs. 

Currently, most companies, likely over 95%, are still in the early stages of exploring GenAI. They're running pilots or small-scale experiments, often using free credits provided by hyperscalers. Because of this, GenAI-related costs aren’t yet a major concern. However, as GenAI adoption becomes more mainstream, expenses are expected to rise, and companies will need to manage them more closely.

At CloudKeeper, we see GenAI as an opportunity. Our customer base has grown from around 300 to over 400 in the past year, and we’re expanding the range of tools and solutions we provide. Internally, we use GenAI to increase operational efficiency and enhance product development. We're also helping our clients identify appropriate use cases for GenAI and navigate the associated cost and implementation challenges. While GenAI has potential, it’s not universally applicable, and we're guiding customers on where and how to apply it effectively. 

How is the global market responding to AI? Are companies mainly experimenting with it, or have they advanced further?

As of now, many B2B companies have emerged. A large portion of them are essentially wrappers around existing large language models (LLMs), often targeting functions like marketing, sales, or HR. Most of these companies use general-purpose LLMs like those from OpenAI or Anthropic and build use-case-specific layers on top, fine-tuning the models for targeted applications.

This pattern is seeing widespread adoption. On the consumer side, tools like Perplexity, Anthropic Claude, and OpenAI are also being used heavily. Companies like Indigo, Zepto, and Lenskart, some of our current customers, are also adopting these technologies. However, for them, cost is not yet a major concern.

How do your tools address the complexities of cloud cost management?

When we engage with customers, our main objective is to help them reduce their cloud costs. We achieve this through three main strategies. The first is rate optimisation. This involves aggregating cloud usage across multiple customers, which allows us to negotiate better pricing from AWS, similar to how group buying works. We then pass those cost savings to our customers. In addition, we help them make the most of discount programs that cloud providers offer in exchange for usage commitments. We guide customers on the right level of commitment and, in many cases, take on those commitments ourselves.

The second strategy is usage optimisation. We use tools like Tuner and CloudKeeper Lens to provide better visibility into cloud spending and to automate ongoing infrastructure adjustments. The third approach is architecture optimisation. We work with customers to review how their applications are built and identify changes that can reduce infrastructure demands. This may involve rethinking how workloads are structured or how services interact.

These three approaches, rate optimisation, usage optimisation, and architecture optimisation, form the foundation of our cost-saving efforts. We package them into solutions like CloudKeeper EDP Plus, which extends AWS’ Enterprise Discount Program with additional benefits. Tools such as CloudKeeper Auto, Lens, and Tuner support these strategies through automation and analysis.

CloudKeeper recently acquired WiseOps. What strategic advantages does this acquisition bring to your clients?

We had already developed two software platforms, CloudKeeper Lens and Auto, over the past five years. Building a solution like CloudKeeper Tuner (formerly WiseOps) had been on our roadmap for some time, but our focus remained on improving existing tools.

We consistently discussed the need to offer automated infrastructure optimisation features. WiseOps already had those capabilities, along with a strong team and a good product-market fit. However, they were acquiring customers independently, while CloudKeeper already had a customer base of 400 that could immediately benefit from such features.

We saw a strategic fit and decided to acquire WiseOps’ technology and team. They’re now fully integrated into the CloudKeeper suite.

Customer feedback has been positive. Many recognise it as the missing piece. We’ve also seen faster sales cycles, with prospects appreciating our end-to-end cloud cost optimisation offering, covering rate, usage, and architecture optimisation. This has significantly strengthened our value proposition.

How do you assist organisations in achieving scalability without compromising on cost efficiency and environmental considerations?

One way we contribute to reducing carbon emissions is by optimising resource usage. For example, if a customer can run their workloads on 80 servers instead of 100, that directly lowers their cloud consumption and infrastructure cost. This naturally supports sustainability goals.

However, in our conversations with stakeholders, typically CTOs, heads of DevOps, or cloud teams, sustainability is rarely a top priority. It may be discussed at the board level, but among those responsible for day-to-day operations, the focus is usually on performance, reliability, and cost.

That said, we do highlight environmentally friendlier options when available. For instance, if a hyperscaler offers services with a lower carbon footprint, we ensure our customers are aware of those alternatives. Still, in practice, sustainability isn't a major factor in most decision-making processes.

Cloud adoption is growing fast. Where do you see the industry heading, and what trends do you expect in the next three to five years?

As more data center leases expire, workloads are steadily moving to the cloud. This trend has been ongoing for the past decade and is expected to accelerate in the coming years.

Currently, GenAI does not contribute significantly to our overall revenue, which exceeds $200 million. However, we anticipate that GenAI will begin to play a more substantial role over time. Along with this growth, we also expect associated costs to rise.

As cloud costs increase, enterprises are becoming more focused on financial operations (FinOps) and cloud cost optimisation. Although these services exist today, the provider landscape is still fragmented. There are over 50 companies globally offering various cloud optimisation solutions.

We expect this space to consolidate over the next three to five years. Like other industries, a few dominant players will likely emerge, while many smaller firms will be acquired.