Companies turn to AI to avoid ‘cloud sprawl’

Companies are turning to artificial intelligence to root out savings in runaway cloud-computing bills, tapping software designed to pinpoint overlapping cloud applications, excess data storage and other inefficiencies across information-technology systems, corporate technology chiefs and industry analysts say. 

The efforts come as cloud-based tools take over an ever-wider range of operations. That, and the murky economic outlook, is prompting many chief information officers and other enterprise technology leaders to take a closer look at cloud costs—even as they increase overall spending.

One key advantage of third-party cloud systems, such as Amazon.com Inc.’s Amazon Web Services or Microsoft Corp.’s Azure, is that they enable companies to quickly scale up or down computing power as needed—rather than paying to run half-empty in-house data centers in order to handle an occasional surge in demand. 

AMAZON ANNOUNCES NEW CUSTOMIZABLE AI MODELS FOR CLOUD CUSTOMERS

But like most online subscriptions, losing track of a tangled web of cloud services can be just as wasteful, said Andy Sealock, a senior partner at West Monroe, a digital-services consulting firm. 

"One of the best things about the cloud is also one of the worst things about the cloud," Mr. Sealock said. "Automation is necessary for the myriad and seemingly infinite number of variables introduced into the cloud-cost equation." Tracking cloud use, he said, long ago surpassed what even the most knowledgeable cloud engineer could possibly do with Excel.

Among other approaches, artificial intelligence is being used to identify cloud-use patterns on a massive scale across an organization by tagging and tracking different apps and data. It can then flag areas of duplication, waste or mismatches between actual usage and existing payment plans. Mismatches can include paying for a higher tier of storage than a given app requires, or using a costly high-performance server where a general-purpose server will do. 

MICROSOFT IS DEVELOPING ITS OWN AI CHIP: REPORT

By integrating cloud vendors’ billing systems into optimization tools, AI algorithms can also find the most cost-effective services for a range of cloud-computing needs, while forecasting demand and comparing future costs under various purchasing scenarios.

Jake Hoffman, chief technology officer at Gnosis Freight, a Charleston, S.C., company that makes shipping-container tracking software for logistics firms, said cloud-pricing structures are typically laid out in terms that require cloud expertise "in order to even understand how much a use case may cost," he said.

AMAZON CEO JASSY STRIKES CAUTIOUS TONE IN LETTER, PREDICTS GENERATIVE AI WILL ‘TRANSFORM’ CUSTOMER EXPERIENCE

Gnosis Freight uses multiple AWS cloud services to retrieve and organize data from sources around the world, requiring complex, interconnected cloud systems that can easily result in cost mismatches, Mr. Hoffman said. To cope, he said, the company has started using an AI-powered optimization tool by New York-based startup Antimetal, designed to help adjust the company’s cloud spending plans and meet expected future demand, based on past use patterns. 

So far, the tool has helped save the company roughly 15% on its cloud bills, Mr. Hoffman said.

"Put simply, cloud systems are incredibly complicated," said Erik Peterson, chief technology officer at CloudZero, a Boston-based cloud-cost optimization startup. He said cloud cost and usage data provided by AWS, for instance, is typically measured in millions of records a day and hundreds of billions every month. "It’s humanly impossible to audit a system of that complexity manually," Mr. Peterson said.

Apptio Inc., a veteran software services firm, has trained AI algorithms to classify cloud use and billing data from market-leading providers, including AWS, Azure and Alphabet Inc.’s Google Cloud, into types of services, apps and other categories, said Jeremy Ung, Apptio’s chief technology officer. 

THESE JOBS ARE SAFE FROM THE AI REVOLUTION – FOR NOW

"This allows our customers to rapidly see opportunities for spend optimization that they may not have otherwise seen," such as duplicated, unused or underused services, Mr. Ung said. The tool also recommends pricing plans based on how a company uses cloud services, he said.

"The resounding sentiment we hear from CIOs and CTOs is that cloud spend has been their company’s fastest-growing expenditure," said Antimetal Chief Executive Matthew Parkhurst. Since launching last year, he said, the startup’s cloud-spend under management has grown more than 500% month-over-month.

Sanjay Jain, chief business transformation officer at WNS (Holdings) Ltd., a global business-process management firm, said the effectiveness of AI cloud-cost tools depends on the "quality of data, the sophistication of algorithms and the expertise of the individuals interpreting the results." Artificial intelligence also requires time to understand long-term consumption patterns, making it difficult to assess costs for new applications deployed with larger cloud ecosystems, he said. 

US NAVY TO USE UNMANNED, AI-DRIVEN SHIPS TO COUNTER SMUGGLING, ILLEGAL FISHING

And though AI tools can help prevent businesses from being overprovisioned in specific cloud services, most companies are unlikely to reduce overall cloud spending given the proliferation of cloud-based apps—not least AI itself, said Dan Tucker, senior vice president at consulting firm Booz Allen Hamilton Holding Corp. The key, he said, is to "avoid cloud sprawl" by spending more than necessary to drive growth. 

Companies worldwide this year are expected to spend a total of $597.3 billion on cloud services, up 21.7% from $491 billion in 2022, according to the latest forecast by IT research and consulting firm Gartner Inc. In part, the gains are being driven by ramped-up spending in recent months on ChatGPT-like generative AI technology, which requires massive amounts of cloud-computing power, Gartner said.

GET FOX BUSINESS ON THE GO BY CLICKING HERE

"CIOs and CTOs today are under tremendous pressure to control costs without sacrificing their need to transform to a digital business and operational model," said Sid Nag, a vice president in Gartner’s technology and service provider group, focused on cloud services. "They are looking for transparency, predictability and efficiency in their expense outlays," Mr. Nag said.