Between August and November of 2025, the share of enterprise code written by AI jumped from 50% to as high as 90%. These findings by theCUBE Research raised big concerns over who builds software and who bears responsibility when it fails. That acceleration continues unabated, as 500 to 1,000 AI-based production workloads are expected to emerge at the edge over the next two years.
Line-of-businesses are increasingly building applications alongside professional developers, and this can make things complicated for IT teams, according to Paul Nashawaty, practice lead and principal analyst at theCUBE Research and ECI Research.
"This is not technical people doing this," Nashawaty told The Forecast in April at the 2026 .NEXT event in Chicago.
"This is people who run retail or finance or whatever at the edge locations, and they're building applications because AI allows for these applications to be created with natural language. You don't have to know how to code."
He sees the enterprise software boosting productivity but the guardrails have not kept pace. Concerns about these capabilities are colliding with an expanding web of regulatory mandates, a fragmented multicloud infrastructure and a workforce that increasingly skews toward generalists over specialists, and Nashawaty sees pressure mounting on CIOs and their IT teams.
Cloud Native to AI Native
While many waves of innovation are feeding into this frenzy, Nashawaty is keen on cloud native technologies that he says are morphing into AI native capabilities. It was described in SiliconANGLE's coverage of 2026 KubeCon, which reported on how IT teams are evolving their infrastructure layer to handle containerized microservices, AI inference pipelines, and autonomous agents that need to run at enterprise scale.
"Cloud native was a rapid growth. The introduction of AI was cloud native on steroids,” Nashawaty said in an interview with The Forecast recorded “It really accelerated the development of applications."
Nashawaty has more than 25 years of experience working in the tech industry. He said IT organizations are vigorously experimenting with new AI capabilities but must keep in mind that they are fully responsible for AI-generated outputs. He reminds IT pros that compliance and governance are as urgent as the innovation itself.
"You're still accountable for what is created by that AI,” he said. “So if you put something out the door that doesn't meet compliance or release information that's not regulated, that's your responsibility," Nashawaty said.
Cloud Native Application Portability
The compliance pressure intersects with a broader infrastructure challenge. With application portability rated as critical or very important by 87% of respondents in Nashawaty's study, the ability to move workloads across clouds without locking into a specific technology stack has become a business requirement rather than a technical preference. He said 94% of organizations now use two or more clouds. Cloud native architectures are at the center of their strategy.
"Application developers are full circle back to the cloud native comment," Nashawaty said. "They're making these applications basically harmonized across these multiple clouds. So you can use any cloud without being bound to a specific tech stack. You can use any cloud to deploy your application anywhere."
A SiliconANGLE analysis of the cloud native ecosystem found that 66% of organizations running generative AI inference now do so on Kubernetes, the orchestration platform at the heart of cloud native architecture, and that 98% of enterprises use cloud native technologies in some form.
It has created a push-and-pull between IT teams and the vendors they rely on, especially as they continue to hire more generalists than specialists.
"What we find is these organizations are pushing back on vendors to reduce the complexity in order to make it easier for them to do their jobs," Nashawaty said.
The broader arc of enterprise application development, from legacy and containerized workloads to AI-generated code deployed at scale by non-technical users, points toward a future where infrastructure must be as adaptive as the people building on top of it.
"It's going to be less about a bag of bits and more about delivering a full solution that's frictionless for organizations to excel."
Rather than modernization, IT teams need to be in continuous reinvention mode.
"App modernization is a treadmill you continuously run," he said.
Video transcript:
Paul Nashawaty: Cloud native was a rapid growth. The introduction of AI was cloud native on steroids. It really accelerated the development of applications. What we're seeing is that developers are no longer just your professional developers. We see builders, lines of businesses being developers, but this creates a lot of challenges. We see a lot of applications being developed at the lines of business, from marketing, from sales, from finance, whatever it may be. What we find in our research is 500 to a thousand AI-based production workloads are being developed in the next two years as applications at the edge locations alone. This is not technical people doing this. This is people that run retail or finance or whatever at the edge locations, right? Manufacturing and they're building applications, but that's because AI allows for these applications to be created with natural language. You don't have to know how to code.In August 2025, 50% of code was being written by AI. We reran the study in November and December and we found that that number jumped from 70 to 90%. So now we have 90% of code being written by AI and that means that the production of applications is accelerating. Now again, it doesn't mean that it's perfectly written code. It just means that it helps you with your operational efficiencies. My role covers past, present, and future states. So a lot of what I talk about is the heritage environments that are being migrated to current state. And that current state, I view as containerized environments, so using containerization from maybe from a VM or from a siloed environment and then looking at futures like things like WASM or web assembly or different ways of using the code in the future. So a lot of this is happening in real time.
There's a lot of migration from heritage environments to new environments for containerization, but sometimes the juice isn't worth the squeeze. Sometimes it's not really worth refactoring those applications and you want to encapsulate those heritage applications and then build new systems of engagement to access those new systems or those existing system of record. There is a kind of colliding of two worlds going on. What organizations need to understand is the acceleration of using AI gives operational efficiencies. However, you're still accountable for what is created by that AI. So if you put something out the door that doesn't meet compliance or release information that's not regulated, that's your responsibility. So that is something that compliance, governance and regulation needs to really come into effect. Where is this all going? This is incredibly important time that we're in. We find that 67% of organizations are hiring generalists over specialists.
That means that skill gap and friction or complexity needs to be removed. So what we find is these organizations are pushing back on vendors to reduce the complexity in order to make it easier for them to do their jobs. So where this is going and where this is moving forward is it's going to be less about a bag of bits and more about delivering a full solution that's frictionless for organizations to excel.
Related:
After AI Boom Comes the Inferencing Economy
Ecosystem Scorecards Help CIOs Avoid Vendor Lock-In
Clouds With Borders: IT Teams Design for Geopatriation
Tension Mounts Between Supply Chain Challenges and AI Adoption
Ken Kaplan is Editor in Chief for The Forecast by Nutanix. Find him on X @kenekaplan and LinkedIn.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
