Multicloud, or hybrid if you prefer, has been the norm for years. That part of the story is settled. What is still very much in motion is how organizations distribute applications across those environments and why those shifts matter operationally.
Our latest research shows a portfolio that is nearly evenly split between modern and traditional applications. On average, 53% of enterprise apps are now modern, while 47% remain traditional. That balance alone should put to rest the idea that legacy has disappeared. It has not. It has simply been repositioned.
The more interesting signal is not the mix, but the movement inside it.
The quiet shift away from traditional on-prem
When comparing where applications are deployed today versus where organizations say they want them to be, one change stands out.
Organizations plan to drop on-premises traditional deployments from 18.0% today to an ideal of 15.2%.
That delta does not vanish into the cloud wholesale. Instead, it reappears primarily as growth in on-premises cloud, which rises from 17.6% today to an ideal of 18.4%. Public cloud and SaaS also increase modestly, but the most telling transition happens inside the data center itself.
This reflects a deliberate modernization path. Organizations are not abandoning on-premises environments. They are refactoring how those environments operate, shifting workloads from static, infrastructure-bound platforms to cloud-like operational models that support automation, APIs, and faster change.
Colocation and edge remain relatively stable. That stability reinforces the point. This is not a wholesale relocation strategy. It is an architectural evolution.
Modernization changes the app delivery equation
As modern application percentages grow, the assumptions that once made app delivery and security “easy” no longer hold.
In the past, consuming cloud-based services often meant inheriting delivery and security capabilities by default. Load balancing, TLS, DDoS protection, and basic access controls came bundled. That convenience masked complexity rather than removing it.
As organizations spread applications across SaaS, public cloud, on-prem cloud, and remaining traditional environments, that implicit simplicity disappears. Each environment comes with different controls, different visibility, and different policy models.
The result is operational friction.
Teams now manage multiple control planes, inconsistent security postures, and delivery behaviors that vary by location rather than by intent. This is where the real cost of hybrid shows up, not in infrastructure spend, but in operational inconsistency.
Platform thinking restores leverage
The organizations that navigate this shift successfully tend to converge on the same approach. They stop treating application delivery and security as environment-specific features and start looking at it as a platform.
A platform approach does not eliminate environmental differences, but it abstracts them. Policies are defined once and enforced everywhere. Traffic management follows the application, not the infrastructure. Security posture becomes consistent even when execution differs.
This restores many of the benefits that early cloud adoption promised. Speed, consistency, and reduced operational burden return, without requiring applications to live in a single place.
More importantly, a platform approach aligns delivery and security with how portfolios actually look today. Distributed, mixed, and constantly in motion.
The reality of operating hybrid estates
Hybrid is not a transitional phase. It is the operating reality.
The data makes that clear. Workloads continue to move, especially from traditional on-premises environments toward cloud-like models. Modern applications continue to grow, but traditional ones persist and will for years.
The real question is whether application delivery and security evolve at the same pace.
Organizations that rely on environment-specific tooling will continue to feel friction as portfolios shift. Those that invest in a unified platform gain consistency, control, and the ability to move workloads without rewriting operational playbooks.
The distribution of applications is reaching equilibrium. The operational model supporting them needs to do the same.
About the Author
Related Blog Posts
From packets to prompts: Inference adds a new layer to the stack
Inference is not training. It is not experimentation. It is not a data science exercise. Inference is production runtime behavior, and it behaves like an application tier.
Compression isn’t about speed anymore, it’s about the cost of thinking
In the AI era, compression reduces the cost of thinking—not just bandwidth. Learn how prompt, output, and model compression control expenses in AI inference.
The efficiency trap: tokens, TOON, and the real availability question
Token efficiency in AI is trending, but at what cost? Explore the balance of performance, reliability, and correctness in formats like TOON and natural-language templates.
Programmability is the only way control survives at AI scale
Learn why data and control plane programmability are crucial for scaling and securing AI-driven systems, ensuring real-time control in a fast-changing environment.
The top five tech trends to watch in 2026
Explore the top tech trends of 2026, where inference dominates AI, from cost centers and edge deployment to governance, IaaS, and agentic AI interaction loops.
