FinOps

A yearslong push to move enterprise IT from on-premises computing to public cloud has delivered more agility and better performance. But for most businesses, modernization also ushered in towering cloud costs. 

The IT spending category soared as technology teams migrated workloads to hyperscale providers and leveraged novel services.

The struggle to keep costs at bay is pervasive among CIOs. Nearly 3 in 4 companies ran past their estimated cloud budgets last year, according to Forrester estimates. Though providers have built tools and processes to curb overspending, broader cloud use rationalization remains elusive.

"At a basic level, it is getting easier, but the demands of what we want, what end users want is becoming harder and harder to fulfill," said Tracy Woo, principal analyst at Forrester, during a CIO Dive virtual event in August. 

One major hurdle is a historic lack of visibility into cloud computing costs. Providers and industry groups have targeted efforts toward standardizing cloud bills across different services. The most salient initiative is the FinOps Foundation's FinOps Open Cost and Usage Specification, known as FOCUS, unveiled in November 2023.

Leading hyperscalers have also pressed forward with onboard cost control measures, including new platform features and more efficient chipsets. But CIOs can also bring cloud costs down by relying on automation and paying attention to people processes surrounding compute.

"Visibility is super important. But I think what is equally important is a culture of accountability," said Jenny Lin, global chief architect at Broadridge Financial Solutions, speaking on the same panel.

Automation and culture

Cloud cost is a complex problem. It requires a mix of human and technological solutions.

Lin detailed how Broadridge responded to a common cloud pain point. One of the company's project had overshot its budget — by about $16 million.

"We set a goal that we were going to remove $16 million of cost in four weeks," Lin said. "And no one thought it was possible." Through daily meetings between its infrastructure, application development, product and business teams, the company worked together to rightsize and shut down workloads until it hit the target, Lin said.

The cultural aspect of cloud cost management is a challenge for most businesses, Woo said. It also tends to be the most difficult for tech leaders.

"As an IT person that is brought up through the organization, you went to school, you learned IT and cloud stuff," said Woo. "You didn't learn how to be a therapist. You didn't learn how to be a mediator."

With more workloads landing on cloud, cost control is the top concern for executives, according to Flexera's 13th annual State of the Cloud Report, published in March. Cybersecurity is a close second in the list of enterprise concerns, but cost has topped the list two years in a row. 

Once a culture of accountability has been ingrained into the organization, automation can help businesses keep spending on track, Lin said. Broadridge automated alerts so that the IT team could see if cloud allocation was surpassing specific levels.

In addition to cloud cost control, another pending task for organizations is to convey the value of the technology.

Tech leaders must keep "articulating the value of the cloud to the senior leadership so they continue to understand that it's not about being cheaper: It's opening up things for us that, even if they cost the same, that it's worth doing," Lin said. "We need to keep reminding people, because it's almost like we keep forgetting."

Most software vendors struggle to manage cloud costs, according to a Tuesday report published by CloudZero. The FinOps solutions provider drew on a survey of 700 SaaS companies conducted by research firm Benchmarkit.

Three in 5 software vendors have not implemented a cloud cost management program, the report found. Nearly three-quarters of respondents said cloud represents at least 20% of their total costs of goods sold.

“For SaaS companies, the cloud is typically a top-three budget line item — yet most don’t have rigorous processes for managing their cloud efficiency or understanding the relationship to their unit economics at any level of granularity,” Phil Pergola, CEO of CloudZero, said in the report.

Companies across industries have felt the sting of spiking cloud costs as they adjust to usage-based billing models and grapple with tying cost to value. Cloud-native SaaS vendors are no different, according to the report.

Nine in 10 surveyed companies admitted they can’t trace at least 10% of their total cloud spend to the right source. Of the 39% of respondents with cost management processes in place, just over one quarter have optimized software code and less than half have enabled chargeback processes that deduct cloud spend from departmental budgets.

Linking cost to a specific unit is complicated by the array of disparate cloud services companies consume, as well as the prevalence of shared resources, multiple cloud providers and containerized workloads.

Without a clear picture of who to bill for which service, cloud consumers can’t even begin to optimize usage — a key feature of FinOps cost management practices.

Most software companies lean on hyperscaler discounts, bulk-rate pricing and third-party resellers to economize in cloud, which can lead to overprovisioning. Only one-quarter of respondents said they regularly use cut-rate spot instances to reduce spending.

When technology providers spend more on cloud, costs generally trickle back to enterprise customers by way of price increases.

Nearly 80% of U.S. organizations shouldered software price increases over the past year, according to Forrester. Most executives attribute the bump to novel software features, including AI.

While AI development is one likely culprit for cost increases, cloud is where large language models are trained, tuned and accessed. If SaaS vendors are overpaying for cloud, they can let those costs erode revenues or pass those bills down the food chain by bumping up the price of their products.

Sponsored content

By CloudBolt Software

As a technology leader of twenty-five years, I have worked on many solutions across a variety of sectors. These solutions delivered clear and tangible value to organizations, as all worthwhile technology investments should. However, when it comes to FinOps, the landscape is starkly different. The aspirations are high, but the actual capabilities remain frustratingly underdeveloped. 

Embarrassingly so, even. 

In fact, two significant forms of disconnect present existential crises for FinOps: our ability to demarcate and demonstrate value to the business and a widening divide from the audiences needed to do so.

These systemic issues not only cripple FinOps’ ability to become activated within organizations but also render its contributions invisible to the audiences who matter most— the executives who are betting on FinOps and the internal stakeholders we support.

In this article, I’ll explore these existential crises and explore solutions that will lead us on the path to up-leveling FinOps’ to its full potential. 

Existential crisis #1: A disconnection to business outcomes

In all my conversations with organizations, a few recurring themes emerge: very few can draw a direct line between their FinOps activities and the actual business outcomes that matter—like profitability, growth, and strategic impact. While the underlying philosophies and intentions in FinOps are sound and impactful, the reality is that most teams are working with KPIs that don’t resonate with C-suite priorities. 

How many FinOps teams do you know that can successfully demonstrate how their cloud initiatives directly contribute to business growth? 

I’ll wait.  

Instead, the output is dominated by primitive signals: raw usage metrics, detailed breakdowns of cloud spend, and operational data that hold little value for CFOs, CTOs, or CIOs. These leaders are looking for insights tied to high-level strategic outcomes, not granular technical data. This gap fosters skepticism about FinOps’ intrinsic value—and honestly, who can blame them? 

Existential crisis #2: A disconnection to engineering activities

In nearly every organization we work with, FinOps teams either overstep or understep when it comes to working with engineering. 

Overstepping often involves an increasing demand for engineers to address every granular data point and action at every step to capture even the tiniest savings opportunity. Understepping is a much lazier version of this, where tickets are auto-created with minimal direction or insight—something I often refer to as the creation of "Jira hell.”

The situation worsens when engineers try to act on FinOps recommendations. They’re often forced to step outside their usual workflows, creating unnecessary friction in an otherwise well-defined and automated process. This disrupts their efficiency and makes working with FinOps feel more like a burden than a benefit. 

It is no wonder that engineers see FinOps as disconnected from reality (an example of this is well-expressed in the following Reddit post).

Much must be done to ensure the relationship between FinOps and engineering is more practical and leads to meaningful action. 

Bridging the gap with third-generation solutions

Third-generation FinOps tools represent the next evolutionary step, breaking down both the horizontal barrier between FinOps and engineering teams and the vertical barrier between FinOps and business outcomes. 

These solutions meet engineers where they are. Unlike second-generation models that overwhelm engineers with overly precise data points, third-generation tools deliver actionable insights by focusing on trends and categories—abstracting data in a way that frees engineers from the noise and allows them to take meaningful action. This shift dramatically improves operational efficiency and drives a higher Return on Effort (ROE). 

What’s more, third-generation tools integrate seamlessly into engineers’ native workflows. This eliminates friction, enabling engineers to manage FinOps tasks without interrupting their usual processes. With AI and machine learning baked in, these tools further reduce manual effort over time by learning from user behavior, making the entire process smarter and more efficient. 

Once FinOps and engineering are aligned, connecting FinOps to business outcomes becomes much easier. Third-generation tools abstract data for engineers and elevate it for C-suite executives, focusing on metrics like total cloud cost savings, unit economics, and cloud spend forecasts versus actuals. This allows leadership to tie FinOps directly to the company’s long-term goals, transforming it into a strategic, revenue-generating function—where it should have been all along. 

The way forward for FinOps

Right now, FinOps is at a crossroads. It can’t move farther with rudimentary tooling. The future of FinOps lies in third-generation solutions that solve these critical challenges, allowing FinOps to finally bridge the gap between engineering and business outcomes. This connection is essential for FinOps to succeed, and by extension, for the entire organization to thrive.

I dive deeper into these third-generation solutions in a recent discussion with Tracy Woo, Principal Analyst at Forrester Research, during our recent webinar, FinOps Reimagined: AI, Automation, and the Rise of 3rd Generation Tools. Watch the recording here to learn how FinOps can finally deliver on its promise.

As tech leaders look ahead to next year, most anticipate IT budgets to increase, according to Forrester. The analyst firm surveyed 2,200 business and technology decision-makers for two August reports.

Nine in 10 respondents expect a bump in IT spending in 2025. Three-quarters of executives anticipate larger budgets for software and technical personnel, driven by growing enterprise appetite for AI capabilities and broader modernization initiatives, the reports found.

Forrester projects tech spend to increase 5.3% year over year in 2024, up from a 4% bump in 2023. But for many organizations, the budget increases will merely keep pace with inflation. “Global inflation is averaging 5.9% and will force tech executives to strategically reallocate IT budgets,” Forrester VP and Research Director Matthew Guarini said in a recent blog post.

Business leaders want more and better technology to streamline operations, improve efficiency and drive the innovation that grows revenue. But in an election year marked by global economic volatility, they can’t risk runaway budgets.

CIOs have grown accustomed to the push-and-pull dynamic underlying digital transformation aspirations. More cloud use led many companies to feel the weight of growing costs. Generative AI’s hunger for high-capacity compute resources is having a similar impact as investments in the technology fail to produce measurable returns.

Cloud and AI will nevertheless be major areas of investment in the coming year, Forrester VP and Research Director Paul McKay said in a blog accompanying the reports. “Organizations are investing even more heavily in cloud, both for AI infrastructure and services support, along with the migration of more of their workloads to the cloud,” McKay noted.

Looking to generative AI for immediate gains is a short-sighted mistake, according to Forrester. While nearly 9 in 10 respondents expect to increase spending on AI-related data infrastructure, talent and tools in the next 12 months, that money will be wasted in the absence of a broad and flexible long-term strategy.

“You’ll need the ability to run AI workloads wherever needed, from the PC to the data center to the cloud, and you’ll need a modern data platform,” Forrester said.

McKay recommends tech leaders increase or defend current levels of enterprise investment in cloud usage optimization practices, trimming spend in targeted areas in order to attenuate the cost of AI workloads.

To offset growing spend on AI, cloud and tech talent, CIOs can reach for a number of other potential cost-cutting solutions.

As platform-based strategies mature, enterprises should focus on eliminating technical debt tied to siloed, single-task applications.

“Years of teams building and maintaining their own applications or insisting on idiosyncratic ‘ideal’ architectures has resulted in massive technology sprawl,” Forrester Chief Research Officer Sharyn Leaver said in a blog post.

“It’s time to inventory your bespoke applications and isolated infrastructure that serves just one or a few applications and replace them,” said Leaver.

The Fintech Open Source Foundation, known as FINOS, published the first draft of an AI adoption framework for financial institutions. The document details common large language model risks and governance measures to guide successful deployments.

“The main obstacle to adoption of GenAI in financial services is the lack of a compliance framework,” FINOS Executive Director Gabriele Columbro told CIO Dive. “We absolutely need a shared standard interpretation of what AI readiness means.”

FINOS, a nonprofit banking and technology industry group allied with the Linux Foundation, welcomed chipmaker Nvidia, credit rating agency Moody’s and model builder Protect AI to its roster of members. Capital One, Citi, Goldman Sachs, JPMorgan Chase and Morgan Stanley are FINOS members, as are AWS, Microsoft, Google Cloud and several other vendors.

While the data-driven financial industry is eager to implement LLM technologies, the path forward is beset with potential risks stemming from AI inaccuracies, data vulnerabilities and an unsettled regulatory environment.

The National Institute of Standards and Technology released a draft of its generative AI risk management framework in July and the European Union’s AI Act took effect in August. But as states like California wrestle publicly with regulating the technology, executives contend with worrisome opacity around best practices and legal liabilities.

FINOS is aiming to bring the same level of scrutiny to generative AI adoption as it did with cloud. The organization launched a Citi-led initiative to fashion public cloud compliance, resilience and security control standards across the industry last year and stood up the AI readiness working group in April.

“This is a great starting point for us, as an industry, to collaborate on a structured approach to the adoption and governance of AI,” Madhu Coimbatore, head of AI development platforms at Morgan Stanley, said in the Tuesday announcement.

The working group limited the scope of the draft framework to internal use of pre-built generative AI tools that are fed additional data through retrieval augmented generation and linked to external SaaS endpoints. The document delineates 16 control procedures to limit 14 specific threats, including data leakage, ineffective data encryption and model hallucinations.

Columbro said the group still has to work through confusion surrounding how open-source methodology applies to LLM technologies.

“Nobody knows what open-source AI means,” Columbro said, noting ambiguity has led to “some degree of open-source washing, where people use the term too loosely.”

As work on the AI project moves forward, FINOS expects to have open-source control code ready for testing and posted to GitHub in the coming months.

Enterprises are expanding mainframe capacity even as they migrate workloads to the cloud, according to an Information Services Group report published in October.

One-third of enterprises using mainframes plan to upgrade processors or storage and 1 in 6 intend to purchase new units in the next three years, ISG found. The research and advisory firm surveyed 200 CIOs, CTOs and other technology executives responsible for on-prem mainframe and mainframe managed services at G2000 companies.

“Though public cloud adoption continues to grow, it will not totally eclipse mainframes in enterprise IT,” ISG Distinguished Analyst Alex Bakker said. “In fact, mainframe compute and storage requirements are expected to rise, particularly in the form of MFaaS.”

As digital transformation strategies mature, cloud cost and security concerns have rekindled enterprise appreciation for the mainframe. Companies with mainframes continue to selectively migrate workloads to the cloud while also shoring up on-prem systems and adding service-based mainframe capabilities.

IBM Z series mainframes remain the machine of choice for more than 90% of enterprises and 59% of respondents are running the z16 line, introduced in 2022, the survey found. IBM and Accenture are the most commonly used vendors for enterprise mainframe services, or MFaaS.

The impetus to modernize has IT executives juggling investments in new technologies with the need to maintain and upgrade existing systems, according to the report. Nearly 60% of respondents said they migrated mainframe applications to cloud in the last three years while more than 40% opted for MFaaS solutions.

“Mainframe demand is being fueled by enterprises’ inability to decommission their legacy infrastructure, often due to conflicting enterprise demands such as the adoption of AI,” Michael Dornan, ISG principal analyst, said. Competing priorities and mainframe challenges will drive enterprises to third-party modernization services to “get the full value from existing investments,” he said.

While enterprises are piloting generative AI use cases in the cloud, many are planning to run workloads on-prem, according to a Kyndryl study published in September. To support compute-hungry large language model tools, IBM will roll out high-capacity processors for mainframe and LinuxOne customers next year.

IBM saw mainframe revenues grow 6% year over year during the three-month period ending June 30. IBM SVP and CFO James Kavanaugh noted the z16 had outperformed expectations, during a July earnings call.

“The picture for mainframe is in some ways rosier than the picture for cloud,” Bakker told CIO Dive in July. “We’re seeing planned growth and spend on mainframe, planned growth and use of mainframe and relatively limited planned retirement of mainframe. Meanwhile, the cloud conversation is all about cost optimization.”

IT executives have learned through experience with cloud that some applications are a better fit for on-prem infrastructure and cost less to run on the mainframe.

“There are a lot of non-scaling workloads running on fully depreciated hardware that don’t need to be in cloud,” Bakker said. “If your data center is relatively modern and doesn't have air conditioners from the 1980s using antique refrigerant, you can run those workloads very cheaply.”

Business leaders are pressuring IT organizations to help streamline operations, increase productivity and enhance customer experiences. The challenge for CIOs: keep costs in check throughout the process.

From cloud to AI and cybersecurity, technology chiefs must strike a delicate balance of enabling powerful technologies without breaking the bank. They must invest with care but avoid falling behind on innovation. 

To maintain healthy levels of IT spending, CIOs and other technology chiefs must learn to quickly identify winning use cases, increase visibility over IT costs across the organization and leverage automation to stave off runaway cloud bills.

Here are four tips from technology executives for keeping IT budgets in line.

(Comments below were lightly edited for length and clarity.)

Feroz Merchhiya, CIO at the City of Santa Monica

We need to be realistic in looking at what our needs are and not get enamored by new, shiny things. Look at the assets that you have available, see what they deliver for you, and then identify the real gaps. 

As technologists, we do get attracted by new and emerging technology. While there's nothing wrong in leveraging those, you have to have a defined use case.

Dan Priest, chief AI officer at PwC, US

One pitfall is to seed the teams with investment funds to go experiment with, but not monitor and manage the ROI. You want to be able to trace the wins back to specific AI investments.  

You also want to understand where you're not winning, and where you need to make some adjustments to your investment strategy, your build approach or your innovation strategy.

Jenny Lin, global chief architect at Broadridge Financial Solutions

Automate everything. Automate settings and create the right guardrails, through alerts and through data transparency, so that decision makers across business and tech can actually make a conscious decision. 

Automating the base case is what we really focused on, leveraging FinOps best practices and putting the people and processes around them.

Tracy Woo, principal analyst at Forrester

One low-hanging fruit is setting up cloud cost reports that go out weekly to different executives or personas. It's not turnkey, necessarily: You still have to leverage the tools that you've provided or whatever you decide to build. That's one of the challenges of FinOps: getting everyone involved. 

Part of the question is: how often are your CIO and CFO involved in these conversations, and how much visibility do they have?

AI adoption is contributing to a spike in cloud costs for most enterprises, according to a report by Tangoe. The technology expense management solution provider commissioned Vanson Bourne to survey 500 IT and finance professionals responsible for enterprise cloud cost management.

Enterprise cloud costs rose an average of 30% in the last year, respondents said. Spending on AI applications and generative AI specifically were cited as top drivers for growing cloud spend by half of respondents.

Nearly three-quarters of respondents said the AI boom has made cloud bills “unmanageable,” the report found. “The cloud’s hidden costs and unpredictable invoices can become the silent killer of GenAI,” said Chris Ortbals, chief product officer at Tangoe.

Hyperscalers spurred enterprise adoption of generative AI and IT budgets are feeling the crunch. As companies double down on hybrid cloud to power large language model technologies, they are struggling to keep cloud bills in check.

“The price-per-unit for what you need to run AI is higher,” Ortbals told CIO Dive. “You're running GPUs and TPUs — it’s not just a typical EC2 unit on AWS.”

More than 9 in 10 respondents reported climbing cloud bills in the last year, due in part to AI applications and automation tools. Roughly the same proportion said their FinOps cloud expense management practices need to include AI and other software services in addition to infrastructure.

The move to bring SaaS under the FinOps umbrella got a boost when the FinOps Foundation broadened the scope of the practice to encompass service-based applications and other cloud-based software solutions in February. IT asset managers joined FinOps efforts at roughly one-third of enterprises to help tame rising software audit costs, according to a midyear Flexera report.

“Our cloud spend is going up because we've added automation and AI,” Ortbals said of Tangoe.

“We're running our own internal GPT, which is a cost we're incurring in the public cloud on top of our normal spend to power our platforms,” said Orbals. “I think a lot of companies are going through that.”

AI has also upped the ante on shadow IT, a prime target area for FinOps-driven cost optimization.

Software procurements that escape the watchful eye of IT comprise more than one-third of SaaS costs, respondents to Tangoe's study said. Two-thirds of companies uncovered wasted spend in productivity software, a category vendors like Microsoft have infused with LLM agents, coding assistants and other generative AI tools.

The trend amplifies the budgetary implications of shadow IT.

“Consumption-based billing just makes it easier to have rogue IT,” Ortbals said. “Departments can just put it on their purchasing card and off they go.”