For decades, modernization has been the siren song of enterprise IT. Tear out the mainframe, move apps to the cloud, declare victory. It sounds bold, and looks decisive on a slide deck. But it regularly turns into a costly and multiyear regret.

There’s a quieter, more practical path that works far more often and comes with a better return on investment (ROI): modernization in place. Keep the core. Upgrade how it works, how it’s integrated, how it’s accessed, and how it evolves to add more value. Instead of spending millions, you save millions, reduce risk, and, most importantly, avoid the costly mistake of rip and replace.

The Mainframe Is Not the Problem

Here’s some verifiable news that sometimes gets lost: Mainframes are here to stay. These foundational, high-volume, highly secure, and economical (more on that shortly) machines power more than 70% of Fortune 500 companies. Banks, airlines, insurers, big tech, and governments strategically run their most critical workloads on them for a reason: they work at scale, under pressure, and with near-zero downtime.

What is obsolete is the way many organizations still view or interact with them. Some believe these powerful computers are stuck in the past, or at least that’s how they view them. Others might think there is only one way to code, manage, or maintain them, when in fact there are a variety of modern, hybrid tools to interface and integrate with them. That’s not a mainframe problem—it’s a perception problem.

Rip and replace assumes the mainframe is the issue. In reality, the mainframe is often the strongest part of the stack; and it’s already a modern platform. It’s less about "mainframe modernization" and more about "application modernization."

The Hidden Costs of Rip-and-Replace Strategies

Big-bang migrations promise clean slates. What they really deliver is risk concentration.

When you replace a mainframe, you’re not just moving code. You’re attempting to re-create decades of embedded business logic — and those weird, no longer understood things the platform does to keep the business running every second. Much of this logic isn’t fully documented. It lives in the system because the system is the documentation.

Projects drag on. Problems surface. Things don’t work as intended. Costs balloon. Deadlines slip. Meanwhile, the old system still has to run, so you’re paying twice (aka a migration tax). Eventually, leadership faces an uncomfortable truth: The new platform isn’t ready, isn’t cheaper, isn’t better, or just isn’t working.

Cue the reversal. We’ve seen this outcome many times before, and the data proves it’s becoming the new industry standard. Analyst firms have been tracking a steady shift for years — reporting that over 71% of organizations have moved workloads back to their core infrastructure. And recent industry surveys suggest that number is climbing even higher, with as many as 83% of enterprises now repatriating workloads back on-premises after realizing the "cloud-first" promise doesn’t match the operational reality.

The “temporary coexistence” of two systems — the proven mainframe alongside a parallel cloud stack — eventually becomes a permanent, expensive weight of redundant double-spending that most organizations can’t afford to sustain.

Money Math Executives Care About

Rip-and-replace projects are up-front, capital-intensive, and speculative in payoff, with organizations spending $3–$45 million per project and budgets overrunning by 287% on average. In short, the financial math of these migrations is often devastating.

To put rip-and-replace reality more bluntly: 3+ years. $45M spend. 287% overruns. 70% failure rates. 5x higher TCO. 0 new ROI. If this is the path you’re looking to for your digital transformation, then “may the odds be ever in your favor.”

For organizations with heavy transactional workloads, moving to the cloud can cost in excess of five times more in total cost of ownership (TCO), even before accounting for the migration expense itself.

Consider the case of a large global bank that recently revealed, "After moving just a small fraction of our [mainframe] workload, the math broke — incremental cloud costs eclipsed our entire mainframe budget. We quickly realized investing in the mainframe is investing in business advantage."

This isn’t just a cost-saving measure, it’s institutionally smarter. By enhancing what already works, your teams can focus on innovation rather than exhausting resources just to rebuild the status quo.

Modernization in Place: The Pragmatic Winner

When it comes to modernizing the mainframe, don’t immediately think “migration.” Modernization in place flips the script. Instead of replacing the engine mid-flight, you improve the cockpit, the instruments, and the fuel efficiency without grounding the plane.

This approach typically includes:

• API-enabling core systems so modern apps can talk to them
• Decoupling user experiences from back-end logic
• Incremental refactoring of high-value components
• Data access and analytics layers that unlock real-time insight
• DevOps and automation with emerging platforms

The result? You get cloud-native development speed and agility with mainframe-grade dependability. And you do it incrementally, which means lower risk and faster ROI. I’m not saying workloads shouldn’t be shifted to the cloud. But they should be cloud-fit apps—not wholesale migrations.

Modernization doesn’t have to be a "burn the ships" endeavor. In fact, the most successful organizations are realizing that modernizing in place is a pay-as-you-go model for innovation. Instead of a risky, multi-year overhaul, you prioritize the components that matter most, delivering tangible value in weeks rather than years.

Industry leaders such as Morgan Stanley have proven this path, tackling complex coding problems by modernizing what works rather than migrating it. This is a successful example of mastering a hybrid strategy.

According to Dr. Howard Rubin, a pioneer in Technology Economics, this level of technology asset class optimization is exactly what distinguishes top-performing companies from their peers across all industries.

These leaders aren't fleeing their core infrastructure; they are mastering a hybrid strategy that embraces a hybrid mix of both cloud and mainframe. They are investing to the tune of 10% or more in their mainframe environments, when compared to their lower-performing industry peers. And nothing scales like mainframe investment. Doubling your footprint can actually result in a 67% cost reduction per unit of compute, according to Rubin’s report.

The fact is, top-performing enterprises are investing strategically in a hybrid future to reduce their cost per transaction, generate more revenue per IT dollar, bring new advances to the market, and transform faster.

Create Value — Don’t Destroy It

The most painful cost of failed modernization isn’t actually the financial investment. It’s lost credibility. Nothing undermines leadership faster than announcing a bold transformation, then watching it unravel in public—or worse, exposing customers to risk in the process, and/or having to quietly walk it back. Outages at Barclays and Santander show what’s at stake with failed payments, locked-out users, and significant customer attrition. And when it goes further, it doesn’t just stumble—it collapses. TSB Bank’s core banking migration failed spectacularly, locking millions of customers out of their accounts, resulting in a staggering $62 million (£48m) regulatory penalty and major operational fallout, with the CEO ultimately stepping down after the migration disaster.

Some transformations don’t create value. They destroy it. Modernization in place avoids this trap. You can stop, pivot, or accelerate based on incremental returns, ensuring that your digital evolution is a series of successes rather than a singular, expensive gamble.

Oh, and by the way—with modern DevOps, open source, and AI now in play for mainframe, agility is no longer a valid argument for migrating away from the mainframe.

The Future Isn’t Either/Or

This isn’t an anti-cloud argument. It’s anti-either/or. As the The Register recently reported, “The unchecked optimism that led to public cloud-first mandates needs a course correction. It's time for cloud-fit strategies.”

“The future belongs to hybrid architectures that respect the strengths of existing systems while embracing modern tools.”

Mainframes don’t need to disappear for an infrastructure to become modern. They need to be unlocked. Rip and replace might sound revolutionary, but modernization in place makes for systems that will actually survive the next decade.

Modernization shouldn't be a gamble where you hope "the odds are in your favor." It should be a series of calculated wins.

Millions saved and overruns avoided.  Business humming, and an intact reputation. Real ROI.

That’s the power of modernization in place. By choosing this path, you swap $45 million in risks and 70% failure rates for incremental ROI. You swap temporary coexistence for hybrid harmony. And most importantly, you ensure that the only thing "ripped and replaced" is a risky migration strategy — not your core business.

Explore how cloud repatriation is reshaping mainframe strategy in Craig Mullins’ “The Cloud Repatriation Trend: What It Means for Mainframe Computing.” Then put those insights to work with the resources available through your SHARE membership at SHARE’d Knowledge.

Michael Dickson is the technical consulting and client services leader for Mainframe DevX and DevOps at Broadcom Mainframe Software. He has more than 20 years of experience leading technology marketing and product management in corporate and startup environments. His expertise includes enterprise systems, software as a service, cloud computing, Linux, artificial intelligence, analytics, and blockchain. Dickson contributes to industry thought-leadership on mainframe modernization, hybrid cloud strategies, and the evolving role of enterprise IT in digital transformation. He also serves as a diversity, equity, and inclusion ambassador and advocates for inclusive practices across the technology industry.