Even as vibe coding lets anyone author apps and compresses weeks of engineering into hours, it is straining the IT systems and people who support enterprise applications.
The surge in vibe coding is moving the bottleneck from developers to the infrastructure teams that must run and scale the resulting apps. Systems operators now face rising cloud costs, staffing conflicts and pressure to consolidate the platforms they use to manage it all, according to Steve Carter, director of product marketing at Nutanix.
"It's so much easier to write code now than it has been in the past," Carter told The Forecast. "It's moving the bottleneck from the engineering team to the infrastructure team."
Vibe coding is spreading quickly across the enterprise. Gartner defines vibe coding as "a fast-paced development style where developers rely on AI suggestions and coding assistants to generate code and rapidly prototype." The analyst firm said it is "a harbinger of the future" that "needs to be taken seriously."
At the 2026 .NEXT event, Paul Nashawaty of theCUBE Research told The Forecast that organizations are experiencing a torrent of line-of-business applications developed using natural language. Carter said he expects to see a similar surge in customer-facing business applications.
According to the 2026 Nutanix Enterprise Cloud Index, 85% of respondents say AI is accelerating their adoption of containers.
These changes have profound implications for the people who plan and architect IT infrastructure. As AI-assisted software development drives up the total number of app deployments, systems operators need to rise to the occasion, Carter said. That means reassessing cloud spending, the platforms they use to scale apps and manage infrastructure, and the composition of the teams who run it all.
"The question is no longer: Do we have enough engineers to build these features?" he said. "It's: Do we have enough infrastructure to run and scale the apps that we've built?"
How Will AI-Assisted Software Development Affect Cloud Strategies?
Vibe coding is a fast-paced software development style in which developers rely on AI suggestions and coding assistants to generate code and rapidly prototype applications, shifting the burden of running and scaling those apps onto IT infrastructure teams.
Systems operators are going back to the drawing board as the rise of AI-generated apps makes their current strategies untenable.
"There's an element of 'We've seen this before,' but never of this magnitude," Carter said.
The coming wave of apps built with AI resembles the explosion in cloud-native apps that transpired in the mid- to late 2010s, only orders of magnitude larger, Carter said. During that time, systems operators largely opted to reduce operational burden by handing off new apps to public cloud hyperscalers. That mentality still exists but is fizzling out.
"The challenge is the public cloud has traditionally been more expensive than on-premises, and that cost premium is only likely to increase with the pace of new apps being delivered," he said.
Carter expects the boom in new apps to inflate public cloud infrastructure costs, outpacing the expenses of operating on premises. He points to ongoing memory chip shortages as an example of what happens to prices when companies compete for the same hardware resources. A similar dynamic could play out for public cloud capacity as a jam of AI-generated apps enters production.
"The companies that can manage these applications cost-effectively have a big competitive advantage," he said. "That's where scalable, easy-to-manage, cost-effective infrastructure is going to be a game changer."
Systems operators are realizing the importance of adopting platforms to do more in their own data centers, Carter said. The shift in mindset is driving interest in Kubernetes, the container orchestration platform that scales applications across local environments, private data centers, the public cloud and edge locations.
"Kubernetes manages a lot of that scalability for you, which is really one of the hardest parts of building apps for this new era," he said.
Who Will Manage the AI Factory?
The boost to app development is forcing managers to make tough decisions about headcount and how to allocate IT staff across competing AI initiatives.
Even as business units pump out more apps for IT teams to onboard, organizations are building out their own AI factories. These specialized, graphics processing unit (GPU)-equipped data centers require the same type of IT practitioners to run and manage them that enterprise apps do, creating a conflict.
"People will probably want or need to move to support the new AI factory," Carter said. "But again, because a lot of the output of AI is more code and more apps, businesses are going to realize that they can't take away from the operations side to move to their AI factories."
The risk is that organizations will stretch crews thin to fill duplicate roles across siloed environments. Companies could end up with a VMware environment for running legacy applications, a Kubernetes environment for running cloud-native apps, and an additional AI environment for running inference workloads, dividing staffing resources and systems operators' attention between them.
"That will be a lot more inefficient than if you can do all of it under the same platform," he said. "You'll still need more people because of increasing infrastructure demand, but you won't need a fully duplicated, siloed team."
Will Vibe Coding Replace Infrastructure Platforms?
With all the chatter about the impact of vibe coding on operations, it is natural to wonder if the orchestration and infrastructure management layers will give way to AI-generated alternatives. The hypothetical SaaS-apocalypse imagines a scenario in which AI agents learn to replicate entire enterprise platforms and systems.
Carter said the complexity of IT systems could likely make it difficult to generate them with AI.
"You would need people who have very deep understanding of distributed systems architecture telling the AI what to build," he said.
The algorithms that underlie the logic in distributed systems are notoriously complex, Carter said, because they must keep many fallible computers in various states of awareness synchronized across a messy network.
"There's a big difference between building apps and architecting complex systems," he said. "And people with knowledge of complex real-world problems are going to continue to be indispensable for solving those problems."
He holds a similar line on the irreplaceability of systems operators.
"That's something AI likely won't replace any time soon," he said. "We need people who can design, plan, manage and operate these systems. Even if AI gets better at it, because the scale is so big, and because these systems are so consequential, people are going to remain central to managing all the complexity."
Related:
Jason Johnson is a contributing writer. He is a longtime content and copywriter for tech and tech-adjacent businesses. Find him on Linkedin.
© 2026 Nutanix, Inc. All rights reserved. For additional information and important legal disclaimers, please go here.
Related Articles
