We live in a multicloud age. According to a recent survey of IT leaders from HostingAdvice.com, 89% of respondents say organizations should never rely on a single cloud provider in their shops. Rather they turn to hyperscale providers like Azure AWS and Google as well as smaller providers for cloud services. But that doesn’t mean everything is smooth sailing. In fact, 73% struggle with security complexity, and 45% worry about vendor lock-in. 

And then there’s a problem looming in the background. How do you make sure your multicloud spending is paying off? What’s multicloud’s return on investment (ROI) — and how can you improve it? For those answers, we turned to executives and cloud professionals to find out more about their tools and tricks of the trade.

How multicloud pays off 

Multicloud architecture offers increased resilience, which is one of the clearest ways they can deliver return on investment. “No one wants to say it out loud, but major cloud outages do happen,” says Bill Hineline, field CTO at Chronosphere. “When they do, having 10 zones in one provider means nothing. That’s when multicloud earns its keep.” For organizations with steep per-minute downtime costs, being able to failover across providers can be a financial and reputational lifeline.

But different cloud providers are more than just identical storage and compute clusters, and a multicloud architecture can provide your enterprise with the best tools for specific jobs. 

“Each of the clouds are better at certain workloads,” says Mike Bilodeau, head of marketing at Baseten. “Depending on your business and what you’re building, it can make sense to just use different clouds for different products.” Hineline agrees: “You might run analytics in GCP [Google Cloud Platform] while keeping core workloads in AWS.” 

Jimmie Lee, CEO of JLEE & Associates, adds that “one cloud provider may offer a lower cost for storage, another may offer better pricing for compute or AI transactions. When implemented correctly, these gains significantly help reduce costs.”

Performance is another place where multicloud can shine. Bilodeau says that locating workloads closer to end users across cloud regions can reduce latency and increase responsiveness. “The classic stat from Amazon is that for every 100ms of added latency, they lost 1% of topline revenue,” he says. “Using multicloud to get workloads closer to customers can help to address this.”

The leverage goes beyond architecture. In AI-heavy environments, inference costs can quickly dominate a company’s cloud spend. “We’ve seen firsthand that for many large companies, inference costs overshadow all other cloud costs by an order of magnitude,” says Amir Haghihat, CTO of Baseten. “Being multicloud gives them non-trivial leverage in contract negotiations.” 

It also increases access to scarce resources like GPUs. “Demand for the latest hardware is so high that getting things like B200s can be difficult,” Haghihat says. “Being multicloud lets you more easily take advantage of the better price/performance wherever those GPUs are available.”

Multicloud also gives enterprises the flexibility to fine-tune their operating cost structure by splitting workloads between public and private cloud. Joe Locandro, global CIO at Rimini Street, saw this firsthand. 

“We were running large predictive models for the energy grid, and that required massive compute power,” Locandro says. “Migrating to a more sophisticated cloud reduced simulation time from 12 hours to two, which allowed the team to run 300% more simulations per month.”

Where multicloud falls short

All those multicloud benefits no doubt sound like a cash-register’s ka-ching to tech leaders. But all the experts we spoke to emphasize that getting those returns on investment depends on execution and the ideal outcomes are often elusive. In practice, many organizations find that multicloud can just as easily introduce friction, redundancy, and diminishing returns.

“Flexibility’s a myth if your team spends half their time context switching,” says Christian Khoury, CEO of EasyAudit.ai. “We saw productivity drop when engineers had to juggle two cloud consoles, two IAM systems, two sets of docs.” Rather than enabling agility, multiple environments can bog teams down with duplicated tooling and incompatible workflows. And once you’ve managed to deploy your applications, you may find that you take a performance hit.

Even the idea of workload portability — a central pillar of many multicloud strategies — rarely plays out the way organizations expect. “One lie that organizations tell themselves is that a multicloud strategy enables applications to move between cloud services as needed,” says Andy Ellis, partner at YL Ventures and former CSO at Akamai. 

“The reality is just about the opposite.” In most cases, apps are tightly integrated with provider-specific services — some of which were the specialized features that led you to choose multicloud in the first place. “You can no longer model applications as a ‘rack of servers, just virtual,’” Ellis says. “Applications are an ecosystem of microservices, many of which you pay for.”

This tight coupling doesn’t just limit portability — it also undercuts negotiating power. “You don’t really have the leverage to threaten to move existing applications easily,” Ellis adds. “Cloud migration projects are arguably more expensive than data center migrations used to be.” Providers know this, and while they have an incentive to retain your business, the cost and complexity of switching becomes a form of de facto lock-in.

Why you need to justify the ROI of multicloud

The right question isn’t whether multicloud saves money — it’s whether it’s worth the cost. That’s how Forrester analyst Tracy Woo frames the issue. “Almost everyone understands that you’re not going to migrate workloads to the cloud just to save money,” Woo says.

 “So the next question is: Is it worth it? And that gets tied into conversations beyond cost management.” Many companies aren’t making these decisions as part of a long-term plan, she says. “A lot of the so-called ‘strategy’ was just: everyone else is doing it, we should do it too. That was especially true during the pandemic.”

For companies still building toward scale, multicloud may not even be relevant — yet. Baseten’s Bilodeau notes that “regarding the latency question, people typically do not realize [they have latency problems] until they hit a certain level of scale.” To solve such problems, “most companies aren’t going to go multicloud at the start, because it’s such an investment in infrastructure versus getting product adoption.” But once a company hits a certain threshold — either in traffic, performance requirements, or customer demand—the pressure to go multicloud grows.

The problem is that this growth rarely unfolds as part of a clean strategic rollout. Often, technical teams default to the tools they know. “We didn’t want multicloud,” says EasyAudit.ai’s Khoury. “We ended up with it because different teams picked what they knew. Devs went GCP. Ops went AWS. Then legal said we needed a private cloud option for one enterprise deal. Welcome to vendor hell.”

Even mature enterprises often inherit multicloud complexity unintentionally. As YL Ventures’ Ellis puts it, “Mergers and acquisitions bring cloud environments with them, and you accumulate cloud services at a rate commensurate with your growth. Development teams may bring expertise in different cloud technologies, and they’ll build with what they know,” he adds. “And partnerships you’ve had in other areas of your business may have driven discounted rates from your partners to adopt their cloud services.”

Sometimes, the decision isn’t even internal. Baseten’s Bilodeau explains that for vendors selling infrastructure software, customer preferences can dictate cloud choices: “Some companies are ‘AWS’ or ‘GCP’ shops. If your software doesn’t run in those clouds, it can be a dealbreaker.” Woo agrees: “A lot of people don’t have a choice. There may have been political reasons, or a rogue engineer, or a partner demanding it — suddenly, you’re supporting what everyone else wants.”

The result is an architecture cobbled together from one-off decisions, rather than a cohesive plan. “Often, we find that our clients are using numerous different cloud services and offerings,” says JLEE & Associates’ Lee. “These were frequently what was architected or recommended to them, without a more comprehensive and multi-year strategy.”

To navigate this complexity, commercial-minded IT leaders need to step up. “Typically, IT departments view cloud providers from a technical and risk lens,” says Rimini Street’s Locandro. “It’s up to the commercially savvy CIO to put in place good metrics, monitoring, and awareness of costs to really understand the value.”

Tips for improving your multicloud ROI: Strategy, structure, and skill

Multicloud can deliver meaningful returns — but only if you’re intentional about how you use it. The benefits don’t materialize just because you have accounts on multiple platforms. 

“When not implemented correctly, multicloud product costs can spiral out of control,” says JLEE & Associates’ Lee. And Rimini Street’s Locandro adds that for many organizations, multicloud TCO, including administration, monitoring, and upskilling workers, “is adding up to much more than initial expectations and vendor promise.” 

Without governance, observability, and architectural discipline, you risk higher costs, team confusion, and missed opportunities. Here are the steps the experts we spoke to recommend to make multicloud worth the investment.

Only use clouds you really need. This one might seem obvious, but it’s a piece of advice often neglected in practice. “Too many leaders see multicloud as a status symbol instead of a strategy,” says Drew Firment, chief cloud strategist at Pluralsight. “It’s not about having ‘more clouds’ — it’s about having the right clouds for the right reasons.” 

At EasyAudit.ai, Khoury’s team asked a blunt question: “What’s the actual value of each cloud in production?” Anything that wasn’t supporting a core service or compliance requirement was flagged. “We weren’t chasing theoretical redundancy—we were trying to kill zombie infra and get one bill under control before explaining three,” he says.

Cloud costs and observability are tied together. “Many teams still treat cost as a line item, not a governance practice,” says Chronosphere’s Hineline. “The same goes for observability says. “Teams that do multicloud well have treated both as first-class citizens. That means tight tagging discipline, clear ownership, and intentional instrumentation. If your FinOps team can’t trace a dollar of spend back to a service and a team, you’re flying blind.”

YL Ventures’ Ellis agrees: “Without visibility into how cloud services are being used, it can be an exercise in driving a bulldozer through a china shop.” Instead of clamping down with heavyweight processes, he suggests “white glove services from the beginning” to onboard teams into cost-efficient environments.

Review your architecture to root out inefficiency. Ellis recommends basic hygiene like identifying duplicate accounts with the same provider. Doing so “can give you better pricing as you purchase in volume,” he says.

Even simple practices can reduce waste. Rimini Street’s Locandro says he’s had “developers spinning up test and dev environments and then forgetting to shut down the environment when they had finished. You need to turn off environments when finished, rather than leave the cloud tap continuously running.”

Measure what matters: Productivity and outcomes. “Multicloud investments need to be measured by meaningful productivity gains and business outcomes,” Firment says. “If you’re not achieving faster deployments and improved customer value, then you’re just racking up cloud bills with a fancier name.”

Upskill your team — and guide them to the finish line. “Flexibility only turns into faster delivery if your people are trained to navigate the differences,” Firment adds. But you shouldn’t expect every developer to master every cloud. “The goal is to build once, run anywhere, and deliver fast.”

That requires architecture that supports smart decisions. As Ellis notes, “Developers will often head for the lowest friction path, so the more you can make it easy for them to adopt the platforms that save you the most money overall, the better off you’re going to be.”

Multicloud complexity demands a smart rollout

If multicloud once seemed like a technical upgrade — or even a badge of sophistication — it now demands something deeper: a strategic lens that connects cloud architecture to business outcomes. That evolution is happening in parallel with the changing role of IT itself.

“I think the context behind [the ROI multicloud] question is something that a lot of teams are being pushed to,” says Forrester’s Tracy Woo. “And the reason they’re being pushed is because IT used to just be IT — you had a desktop, you had some applications loaded on it. Now IT strategy has really become this enabling backbone for business or digital transformation,” she says. “It’s started to level up higher and higher, and that’s why you have cloud really transforming the way people do work. The bigger question that looms beyond that: What’s the ROI for multicloud?”

For organizations that treat multicloud as a checkbox — or let it sprawl without structure — the answer may well be “not much.” But for those that align cloud architecture with business priorities, cost discipline, performance targets, and a realistic assessment of team capacity, multicloud can be the foundation for long-term agility and resilience, and provide a return on the initial necessary investment and then some.