Legacy application code and data are typically much older than the mainframe technology they run on. Mainframe manufacturers like IBM have maintained backward compatibility for decades, ensuring that today’s 17th-generation mainframes can still run 50-year-old applications, as well as the latest containerized cloud-native code, open source applications and middleware.
So, while old software systems may be inflexible and increasingly difficult to support, their mainframe infrastructure is extremely capable and exceptionally secure. Replacing this computing hardware is simply unnecessary — and it can result in modernization projects that stretch a decade or more and cost hundreds of millions or even billions of dollars.
“The mainframe isn’t the problem; it’s the ancient applications and data that have not been updated to take advantage of modern cloud-native services and new processing architectures,” says Jan Gravesen, an IBM Distinguished Engineer and Account Technology Leader for the state of California. “The modern mainframe represents the most powerful and secure transaction-processing and data-serving platform in existence, at a fraction of the power and carbon footprint that a comparable cloud infrastructure requires.”
Hybrid computing — a relatively new processing and modernization architecture — lets organizations leverage cloud-native capabilities while keeping their high-performance mainframes. It’s one reason why installed mainframe capacity worldwide continues to grow, Gravesen says, adding that the technology is a mainstay for banking and finance institutions, defense systems, booking and reservation platforms, and other critical sectors.
“What we have done with hybrid cloud technology is enable the mainframe to behave, look and feel like a modern cloud platform while preserving its massive and elastic performance and impenetrable security,” Gravesen says.
A Better Approach
A hybrid approach that combines the best of the mainframe and the cloud offers a compelling modernization path for California state and local government agencies.
A 2023 California State Auditor’s report found that more than half of surveyed state agencies had at least one important IT system that was close to or at least 20 years old.1 This is not necessarily a problem. Within the past 20 years, new software systems have mainly been developed in Java, .NET, Linux and Windows, running on distributed Linux or Windows servers or on mainframes that themselves are massive Linux servers.
While these systems were created before cloud computing and built on available computing platforms, they largely follow modern design principles and are friendly to change.
California state agencies also use software systems that are much older — in some cases dating to the 1980s. But even these ancient systems often run on modern 15th- and 16th-generation mainframes.
A hybrid approach evaluates existing applications to determine which ones should continue to take advantage of the mainframe’s performance, and which should move to a cloud platform.
When it comes to transaction processing and data serving, the modern mainframe provides qualities that can’t be replicated on cloud platforms at acceptable costs per transaction. Mainframes can process up to 1 million transactions per second — with real-time AI built into the transaction — while being virtually unhackable.2
“No transaction processing system on the planet can beat the bandwidth, security, economics, capacity or throughput of mainframes,” says Venkat Krishnan, Vice President, Senior Partner and AI and Data Transformation Lead at IBM Consulting. “At the same time, people want to migrate to the cloud and derive some of the benefits that accrue from the cloud.”
Hybrid cloud computing also recognizes that modernization decisions must consider factors beyond cost and performance, such as:
- Shrinking Talent Pools: Many original mainframe applications were written in legacy languages like COBOL, PL/I, REXX, JCL or Assembler, which are rarely taught today. The generation of engineers who built and maintained these applications are retired or retiring soon.
- Change-Resistant Programming: Early computing hardware had limited memory, storage and processing power. To maximize performance and minimize cost, developers wrote programs that were tightly integrated and optimized for the underlying data and vice versa. This delivered superior performance, but it made change difficult. Today, the cloud and modern mainframe platforms offer abundant computing power, allowing for looser coupling of applications and infrastructure that’s friendlier to change. But reversing tightly bound data and logic designs require documentation and expertise, both of which are in short supply.
The Technology Behind Hybrid Cloud
Bringing cloud and mainframe together as one — while disentangling decades-old, tightly coupled applications data — may seem daunting. But significant advancements in hybrid cloud technology make this a viable and practical modernization path.
A hybrid cloud strategy involves a two-way exchange of data between mainframe systems and the cloud. Hybrid cloud technology bridges both environments, enabling IT teams to develop and manage mainframe code with the same tools they use for cloud platforms like AWS Cloud or Microsoft Azure. The same development environment can be used on both platforms to create containerized code in modern enterprise languages such as Java, Go and Python. Both platforms can also use the same site reliability and observability practices.
Innovations in mainframe technology coupled with AI and modern application programming interfaces (APIs) enable a cloud-like experience on mainframe systems.
“IBM bet early on Linux, so modern mainframes support open source tools and integrations that simplify connections between mainframes and cloud applications,” Gravesen says. “Now agencies can run cloud-native tools directly on the mainframe, supporting digital transformation without compromising data security.”
Hybrid cloud technology also enables real-time analytics at a lower total cost of ownership compared to other technologies. Modern mainframes use API-driven technologies and software libraries to run popular AI models and open learning frameworks, such as PyTorch and TensorFlow, using built-in, on-chip AI accelerators to process predictions faster. This lets agencies run intelligent applications more efficiently.
Hybrid accelerates AI by allowing agencies to run AI models on the same systems where their production data already lives. This taps into the economic benefits of data gravity — the tendency of large amounts of data to attract other data, applications and services. Agencies can use the latest advancements in AI, including generative and agentic AI, to deliver better service outcomes, and often execute AI models that are embedded in real-time transaction processing.
“No other platform in the world can do this at these speeds. This is why the modern mainframe still executes 90% of all credit card transactions in the world and 72% of all business transactions,” Krishnan says.3,4
The Key Benefits for Government
Hybrid cloud projects are faster and less expensive than mainframe replacements, yielding results in as little as three months, Gravesen says, and at a tenth of the cost.5 In addition, hybrid cloud reduces complexity, enabling agencies to switch or integrate third-party applications easily without getting locked into rigid architectures.
Agencies can use a hybrid cloud to support a range of important use cases.
DevOps: Agencies can execute DevOps workflows directly on the mainframe. Current-generation mainframes also support modern operating systems and coding languages like Python and JavaScript.
AI model development: Modern mainframes handle AI model development and real-time transactional AI inferencing, which underpins activities such as real-time fraud detection. Hybrid cloud also lets agencies avoid moving production data to separate data lakes. They can modernize in place using their current systems.
HHS modernization: Hybrid cloud supports medical enterprise data warehouses and platforms that give health and human services agencies a 360-degree view of client interactions. It also streamlines tasks like verifying eligibility and processing benefits applications.
Digital constituent experience: Hybrid cloud lets agencies deliver omnichannel experiences across various touchpoints, including mobile, chat and the web. It also improves employee productivity by supporting AI-driven automation to streamline workflows and reduce administrative bottlenecks.
Hybrid Cloud in Action
California agencies are already reaping the benefits of hybrid cloud.
The state’s Medicaid modernization project used a mainframe backend and open APIs to create a portal for 260,000 healthcare providers who deliver care to Medicaid and Medicare beneficiaries.
“This is an example where, instead of building that capability on the mainframe, you build it on the cloud,” Gravesen says. “You just integrate it into the existing mainframe using modern open APIs for much faster results.”
The California Department of Motor Vehicles also embraced this approach, bringing together hybrid cloud technologies and modern APIs to perform advanced analytics and to integrate its Salesforce platform, enabling new digital services for residents.
A New Framework for Mainframe Modernization
Keep these recommendations in mind as you consider a hybrid cloud approach.
Understand your options: Don’t automatically think of mainframe replacement as your first, only or best option. IT and business teams should explore new technologies and assess the full spectrum of hybrid cloud options available.
Make a plan: Thoroughly analyze your existing workflows and develop a roadmap for how mainframe modernization will reshape the user experience for employees and constituents.
Partner carefully: IT modernization projects typically require collaboration with third-party technology providers or systems integrators. Some vendors may push agencies toward full mainframe replacement. Look for partners that offer a range of options and innovative approaches to match your modernization and service delivery goals, as well as your overall IT investment strategy.
Focus on change management: Break down barriers between IT teams who maintain the mainframes and those implementing the new technology. Agencies should also align internally on the vision and desired outcomes for hybrid cloud architecture. Appropriately train and set expectations for stakeholders as part of this process.
Establish and track key performance indicators (KPIs): Effective KPIs can include cost reduction metrics, velocity improvements related to process and service delivery efficiencies, application completion rates, and application processing times.
Moving Forward
Mainframes aren’t going anywhere. But, hybrid cloud lets California agencies combine innovative cloud-based technologies with the mainframe’s unrivaled performance, security and reliability.
“The mainframe as critical infrastructure is here to stay,” Krishnan says. “When you want to do very high throughput and secure transaction processing and analytics, there’s no other system on the planet — including the cloud — that can process at that level of capacity and economy.”
1. https://information.auditor.ca.gov/reports/2022-114/index.html
2. https://community.ibm.com/community/user/ibmz-and-linuxone/blogs/meredith-stowell1/2025/02/26/the-modern-mainframe-a-computing-powerhouse?communityKey=e7b7d299-8509-4572-8cf1-c1112684644f
3. https://planetmainframe.com/2022/12/relevance-of-mainframe
4. https://images.sw.broadcom.com/Web/CAInc2/%7Bee65bc93-3ea0-4eb4-bcac-654c682650e3%7D_MSD_Technology-Economics-Report-May2023.pdf
5. https://images.sw.broadcom.com/Web/CAInc2/%7Bee65bc93-3ea0-4eb4-bcac-654c682650e3%7D_MSD_Technology-Economics-Report-May2023.pdf
