How to Navigate the Cloud’s Newest Spending Problem: GenAI

How to Navigate the Cloud’s Newest Spending Problem: GenAI

This article originally appeared in Forbes.

For years, innovation has spurred glutenous cloud spending, but it’s now swelling to new peaks with the introduction of generative AI (GenAI).

Years ago, innovation leaders cautiously nibbled on the cloud with concerns about security and reliability, but large language models (LLMs) can’t be devoured fast enough with GenAI exploding onto the scene. However, feasting on them may shock CIOs and CFOs when cloud invoices strike.

Here’s a closer look at the cloud’s newest spending problem and how to navigate it.

GenAI: A New Kind of Imperative

Pressure is a familiar feeling for IT and product leaders staring down security threats and increased demands for innovation, but the commands for GenAI have shifted from cautious to urgent: “Find the use case for this,” “If we don’t, we won’t be a leader in our space” or even “Invest—even if payback is obscure.”

Consequently, investments are accelerating. According to Forrester, annual spending on specialized applications of GenAI will increase 36% by 2030. Also, theCUBE Research found that roughly 36% of IT decision-makers say their spending on AI will increase by double digits, as of July 2023.

Mandates are manifesting into capital outlay—but at what cost? We’re starting to get a glimpse of the fallout from GenAI.

GenAI Backlash: Most Can’t Feed Its Demands

The competitive advantages of LLMs are being offset by IT and financial blowbacks. “For generative AI, a stubborn fact is that it consumes very large quantities of compute cycles, data storage, network bandwidth, electrical power and air conditioning,” explained Stan Gibson, a writer for CIO.com.

Prerequisites include scalable, high-performance compute power that only modern cloud platforms can supply—it’s a resource hog that IT systems and budgets alike aren’t prepared to feed. Network infrastructure is an AI bottleneck, according to Accenture, which found that 87% of AI systems have outstripped their current network capabilities. To innovate, companies have always had to pay the cloud tax, but GenAI appears to be jacking up the tax rate.

What’s more interesting is the risk tolerance that surrounds it all. Companies are still marching forward regardless, hobbling along with poorly performing platforms or plundering the IT portfolio. “Generative AI is considered a priority for most enterprises, even if it means working with under-optimized infrastructure that they are unwilling or cannot afford to change,” reported David Linthicum, Contributor for InfoWorld.

As pointed out in theCUBE Research article: While 94% of companies state they’re spending more on AI this year, they’re doing so with budget constraints that will steal from other initiatives.

GenAI at any cost can be a risky gamble.

GenAI’s Silent Killers: Cloud Costs and Financial Risks

IT and product leaders are turning to cloud infrastructure services to rapidly deploy GenAI tools, but the cloud is known for twisted economics that, when left unmanaged, can trigger runaway costs and render innovation financially unsustainable.

When the cloud consumes more of the IT budget without generating measurable returns, it can become the silent killer of GenAI.

Twisted Cloud Economics

Scalability is arguably the most famous trait of cloud infrastructure, but that benefit comes with a dark side. When services have no boundaries, neither do invoices. Expenses can be unpredictable. Slight increases in utilization can result in significant cost spikes—sometimes racking up tens of thousands of dollars in hours.

And it’s not just the aftershock of infrastructure as a service (IaaS) invoices. GenAI can quietly foster more shadow IT challenges. Many workers report they are using GenAI tools without the permission or knowledge of their employers. Plus, trends in cloud inflation only multiply AI’s cloud spending problems.

Missing ROI

Invisible cloud costs are the peril of GenAI and the fountainhead of a glut in IT budgets, but what’s more concerning is that GenAI has not yet lived up to its promise for many companies. One study shows 54% of business leaders expect AI and GenAI to deliver cost savings in 2024. While the jury is still out, these may soon become empty promises considering most CIOs struggle to find ROI.

Safeguarding Investments in GenAI

Readying yourself for long-term increases in cloud consumption is a key strategy in safeguarding the success of GenAI innovation. Here are a few steps to consider:

Cloud Costs Shouldn’t Be a GenAI Afterthought 

Perform a risk analysis by evaluating your cloud network’s capacity to meet the performance demands of GenAI. Dig into the details of cloud infrastructure contracts and usage, understanding how effectively you’re using cloud resources and whether you’re leveraging long-term commitments in exchange for lower rates. Evaluate your ability to use money-saving tools like instance pausing across multiple cloud service providers, knowing which pausing schedule and structures will keep more dollars in your pocket.

Know How to Place GenAI Workloads Cost-Effectively

The difference between public and private cloud costs can be significant, making it important to define financial tipping points. Deploying large workloads comes with high stakes, particularly when public costs can spiral out of control and exit strategies can be difficult.

Sharpen ROI Calculations

Executives need to get better at calculating the business value of tech investments. That means tying GenAI expenses to innovation projects and payouts. Financial chargebacks and cost allocation activities can kickstart ROI tracking, establishing any missing links, as I’ve written about in a previous Forbes article.

Finding Funding from within Doesn’t Mean Robbing Peter to Pay Paul

You don’t need to plunder—instead, rightsize. IT expense management platforms, like Tangoe One, can help to track expenditures, identify unused cloud resources and turn IT waste into available funds. Cloud cost optimization programs, led by FinOps best practices, are paramount.

Conclusion

The rapid adoption of GenAI is redefining cloud consumption habits, spurring more spending, outpacing current infrastructure capabilities and threatening long-term success.

While the benefits of GenAI are promising, hidden expenses and unpredictable invoices can undermine investments if not carefully managed. A thorough financial risk analysis, strategic deployments and tighter financial management are essential to ensure that the promise of GenAI translates into sustainable innovation instead of technical debt and cloud budget blowouts.

Ready to ensure your innovation is financially sustainable? Talk to Tangoe.