Your Cloud Economics Pulse For May 2026

Welcome to May’s Cloud Economics Pulse, CloudZero’s monthly look at how cloud spend is shifting as AI becomes a permanent fixture of the enterprise budget.

This month, we lead with what’s happening in the C-suite. Lots of nerves there around AI right now.

According to Dataiku and Harris Poll’s 2026 Global AI Confessions report, which surveyed 900 CEOs at companies with $500 million or more in annual revenue, 80% say their role will be at risk if their company fails to deliver measurable business gains from AI by the end of 2026. 

They don’t mean their team. They’re talking about themselves. 

And it gets spicier: 87% say they would stake their job on delivering measurable results from their AI program.

That sounds like confidence, but there’s a lot of desperation in the boardroom dressed as conviction.

Boards are applying direct pressure (confirmed by 62% of CEOs), confidence in AI execution is slipping (41% last year to 31% this year say they’re “extremely confident”), and 77% outright believe a CEO will walk the plank sometime in 2026 due to a failed AI strategy.

Why now? Because AI spend has crossed a threshold that makes it impossible to ignore.

Listen: AI and ML workloads now consume an average of 5.86% of total cloud spend across CloudZero’s customer base — and the acceleration is very steep in recent months. As recently as November 2025, that number was 2.42%. It more than doubled in five months. It’s now a line item that someone must own, forecast, and justify.

The C-suite pressure and the spend curve are converging at the same moment. In this month’s Pulse, we dig into the usual data of spend by provider, by service, and by AI/ML, and then why CEOs, CFOs, and boards need answers now.

How We’re Looking At Data (And Why It Matters)

For the Cloud Economics Pulse, we track monthly cloud spend trends using anonymized, aggregated data from CloudZero’s network.

Together, these views show not just where cloud dollars go, but how spending patterns shift as new technologies — especially AI — reshape the cost landscape.

One note on methodology: You may find that the AI/ML percentage in the service category section differs from the average in the dedicated AI/ML section. Both are correct; just measured differently. The service category is money-related; essentially, total AI/ML dollars divided by total cloud dollars across all customers. Meanwhile, the average/median figures are org-weighted — every customer counts equally regardless of size. 

Research Report

FinOps In The AI Era: A Critical Recalibration

What 475 executives told us about AI and cloud efficiency.

Main Highlights For May 2026

1. Cost By Provider

Here, we’re looking at how overall cloud spend is distributed across providers: 

AWS still runs the show, but where the rest of the money goes is shifting fast.

AWS has stopped bleeding for now. AWS sits at 69.5% in April, essentially unchanged from January 2025 (69.6%). The 5.7pp erosion from early 2024 is real, however, but it already happened. It isn’t about whether AWS is losing share anymore, rather, who’s taking over that part of the provider pie.

The AI providers are the new entrants — and the pecking order is already shifting. OpenAI peaked at 0.86% in February and slid to 0.59% in April — shedding 0.17pp in a single month. Anthropic moved the other way, gaining 0.06pp to hit 0.27%. Not a coincidence: Deloitte’s 470,000-employee Claude rollout launched in late 2025, Accenture stood up a 30,000-practitioner training program, and Claude Code’s enterprise subscriptions quadrupled since January. Regardless, combined AI provider share is still growing, but the one-vendor era is over.

Marketplace is quietly eating the procurement playbook. AWS Marketplace went from 1.48% to 4% in 15 months — the single largest share gain of any provider in the dataset. GCP Marketplace, meanwhile, swung from 0.35% to 0.82% in a single month — whale-driven and wildly unpredictable. Same label, very different procurement dynamics.

Key takeaways

2. Cost By Service Category

Here, we’re looking at how overall spend is distributed across cloud services:

The cloud spend pie is reshuffling and, for the first time, the direction is clear.

Compute is quietly losing share. Down from 49.5% in January 2025 to 47.3% in April, and shedding 0.64pp just MoM. Still the biggest line item by far, but the trend line isn’t ambiguous anymore.

Storage is the sleeper hit. Crossed 12% for the first time in April, up nearly a full percentage point in a single month and the largest monthly storage jump in our dataset. Storage was relatively flat throughout 2025, then broke out just as AI/ML crossed 2%. That’s no coincidence. Checkpoints, embeddings, vector stores, RAG corpuses, inference logs; all these things mean storage debt on every transaction using AI.

Databases are quietly being unbundled. The category sits at 10.1% — shedding another 0.56pp just this month. This marks an upstream migration to analytics platforms like Databricks (1.27% and climbing) and Snowflake (which plateaued at 1.57%). What we used to call “database costs” is now spread across three or four line items not always being packaged together.

Observability is losing its sacred-cow status. Datadog’s share averaged 2.17% in 2025 and dropped to 1.70% for February through April 2026 — that’s a 24% YoY decline in service pie ownership. When combined with New Relic, it’s a drop from 2.35% to 1.89%. OpenTelemetry is the likely culprit: 48% of organizations now run OTel in production (per Elastic’s survey data), with early adopters reporting 35–67% cost reductions (Grafana). Turns out “you can’t cut what you can’t see” isn’t the conversation-ender it used to be.

Key takeaways

3. Cost Of AI/ML

Here, we’re looking at how AI and machine learning costs are growing as a share of total cloud spend — shown as both average and median percentages to capture the full distribution of adoption across organizations:

This is the chart that gets forwarded to the CFO.

Average AI/ML hit 5.86% in April — up 4× from January 2025. But the acceleration is lopsided: the November-to-April window alone added nearly 3.5 percentage points (3.44%). Just as dramatically, February, March, and April produced the three largest single-month jumps in our dataset’s history — back to back to back (+1.20pp, +0.86pp, +1.02pp). This is a hockey stick with the blade still growing upwards.

The median is the number that should worry you. The typical customer — not the outlier, the median — is now at 1.32%. That’s up sevenfold from 0.18% fifteen months ago. But the real story is that the last three months alone added more share (+0.71pp) than the entire twelve months of 2025 combined (+0.39pp). The average sits at 4.34%, pulled up by heavy adopters, but the gap between median and average has narrowed from 8.6× to 3.3×. In other words, this is no longer a few whales skewing the data. The data is very mainstream and increasingly reflects the broad middle.

Where this goes next is the real conversation. The median doubled in five months (November to April). If it doubles again over the next 12 — and the curve says it will — the typical org lands around 2.5%, with the average north of 8%. At that point, AI overtakes Databases as a cost category for most companies. Let that sink in.

Key takeaways

Analysis: AI spend has crossed the accountability threshold

We can’t ignore this any longer. There’s a moment in every cost category’s lifecycle when it transitions from discretionary experiment to formal budget line item. That’s where finance notices, where the CFO asks “what are we getting for this?”, and where the answer can no longer be “we’re still figuring it out.”

AI just hit that moment. And the consequences are no longer theoretical. Dataiku’s survey found that 65% of CEOs now worry more about over-investing in the wrong AI initiatives than about under-investing and falling behind. The fear has flipped. The question has evolved from “are we spending enough?” to “are we spending on the right things, and can we prove it?”

The spend side: what our data shows

To reprise: across CloudZero’s customer base, the median organization now allocates 1.32% of its total cloud spend to AI and ML workloads. Fifteen months ago, that number was 0.18%. As recently as last November, it was 0.52%. The acceleration is compressed into a window short enough that most annual budgets haven’t caught up. At 0.18%, AI spend disappears into “other” on a cost report. At 1.32% and climbing at the fastest rate in our dataset, it’s a number that someone owns, someone forecasts, and someone has to justify to finance.

Think of it as three zones, with percentage baselines as examples:

The median customer just entered the visible zone. The average customer crossed into accountable territory over a year ago. And the fastest movers, those pulling the average up, are well past the point where AI spend competes with compute for budget attention.

The pressure side: what the C-suite is saying

The CloudZero data shows the spend trajectory. External research shows the consequence of that trajectory.

According to Kyndryl’s Readiness Report, drawing on insights from 3,700 business executives, 61% of CEOs say they are under increasing pressure to show returns on their AI investments compared to a year ago. 

Dataiku’s survey reinforces the pattern: 62% of CEOs say their board is actively applying pressure to deliver measurable AI-driven outcomes, rising to 72% in the U.S. (up from 61% just a year earlier). The pressure is hardening into expectation.

The CFO, likewise, faces accountability. An RGP survey of 200 U.S. finance chiefs (published December 2025) found that 48% of CFOs said they are ultimately responsible for ensuring AI delivers measurable value. Yet only 14% have seen clear, measurable impact from their AI investments to date. The gap between ownership and evidence is stark: those accountable for AI ROI largely can’t demonstrate it yet.

PwC survey results paint a similar picture: only 12% of CEOs say AI has delivered both cost and revenue benefits. Gartner, surveying 782 infrastructure and operations leaders in late 2025, found that only 28% of AI use cases fully succeed and meet ROI expectations, while 20% fail outright.

The organizations that can’t answer the ROI question are already acting. S&P Global data shows that 42% of companies abandoned most of their AI projects in 2025, up from just 17% the year before — often citing cost and unclear value as the primary reasons. Gartner predicts another wave: over 40% of agentic AI projects will be canceled by end of 2027 due to escalating costs, unclear business value, or inadequate risk controls.

The hiding place is gone

This matters because of where AI costs have been living. Constellation Research, reporting on AI budget patterns in 2025, found that companies were deliberately not creating separate AI budgets — instead incorporating AI costs into existing business cases and other line items. The explicit motivation: to continue AI initiatives without committing the kind of dedicated budgets that would require extensive justification and scrutiny. Constellation predicted that “rational AI spending and dedicated budgets will be a 2026 story.”

The CloudZero data confirms the prediction. When AI was 0.18% of the bill at the start of 2025, hiding it in compute or other categories, whether by design or due to provider structure, still worked. With the median and aggregate AI spend rising exponentially, the spend is now too large and too visible to bury. It’s visible. Finance sees it. The CFO sees it.

The question is no longer “how much are we spending on AI?” It’s “what are we getting for it?” — and most organizations don’t have a good answer yet. According to Dataiku’s CEO survey, 56% of CEOs now admit that competitors have deployed AI strategies they consider superior to their own. Confidence isn’t just softening — it’s cracking under scrutiny.

What this means — and what to do about it

That’s the stick. Here’s the carrot: the organizations that can answer the ROI question aren’t just surviving the accountability phase. They’re pulling away from everyone else at a pace that’s getting harder to close.

Grant Thornton’s 2026 AI Impact Survey of 950 business leaders found that organizations with fully integrated AI are nearly four times more likely to report revenue growth than those still piloting (58% vs. 15%). Not cost savings or efficiency gains — definitive revenue growth.

PwC’s AI Performance Study, published in April 2026 and based on interviews with 1,217 senior executives, puts the gap in even starker terms: 74% of AI’s total economic value is being captured by just 20% of companies. The other 80% are splitting the remaining quarter. Those leading companies aren’t just deploying more AI tools — they’re also two to three times more likely to use AI to pursue new revenue opportunities, and they’re redesigning workflows around AI rather than bolting it on. Their employees are twice as likely to trust AI outputs (which means they actually use them). And they’re increasing decisions made without human intervention at 2.8× the rate of their peers — compressing decision cycles in ways that compound quarter over quarter.

The definition of AI success is shifting too. According to Dataiku, revenue growth nearly doubled as the most-cited measure of AI success — rising from 16% in 2025 to 28% in 2026, nearly tied with productivity (25%). The question is no longer “did AI save us money?” It’s “did AI grow the business?”

So let’s add it up. The 20% of companies capturing 74% of AI’s economic value are growing revenue at four times the rate of those still piloting. Their employees trust AI outputs enough to actually use them. They’re compressing decision cycles at nearly triple the rate of peers. And they’re redesigning workflows around AI rather than duct-taping it onto existing processes. The difference isn’t better models or bigger budgets. It’s accountability: the ability to show what AI costs, what it delivers, and who owns the outcome.

For finance teams, that means AI costs need the same intelligence infrastructure that cloud compute got five years ago: visibility into what’s being spent, by whom, and on what at the model, feature, and customer level; unit economics that connect AI spend to business outcomes; and optimization levers that go beyond “spend less” to “spend on the right things.” That’s how the 20% built their advantage. It’s also the only way the other 80% can close the gap before it becomes permanent.

For a deeper look at how global organizations are approaching this challenge with proprietary data and frameworks, download CloudZero’s FinOps in the AI Era report — based on survey findings from 475 senior leaders navigating this exact inflection point.

Your Takeaway For This Month

The experimentation era is over. AI spend is now large enough, fast-growing enough, and visible enough that it demands the same accountability as any other major budget line. The organizations that can connect AI costs to business outcomes will scale their investments and join the 20% capturing the lion’s share of AI’s economic value. The ones that can’t will cut — and many already have.

The spend curve and the accountability curve just collided. Which side of that intersection is your organization on? Thoughts, comments, disagreements? Reply to this Pulse or email [email protected] with “CEP” in the subject heading. We’ll feature the best feedback in an upcoming issue. Watch for our next Cloud Economics Pulse on June 9, 2026, and on the second Tuesday of every month.

Author: Keith MacKenzie

Keith is CloudZero's Content Marketing Manager with a specialty in topics around FinOps, SaaS, AI, and overall work optimization. He also brings more than a decade's worth of background as an editor and writer in the mainstream media and content marketing industries. He's also been to Chernobyl twice – and (sometimes) still has all his marbles.