For a long time, relying on multiple cloud providers seemed the obvious answer to the risks of vendor lock-in and outages. Spreading applications and data across several platforms was expected to offer more flexibility and resilience.
The cloud enters a new phase of control
Yet, each cloud comes with its own tools, security policies, and connection methods. This complexity is amplified by data centers, online software, and legacy systems. Each environment can perform well on its own without necessarily working effectively as part of a coherent whole.
The challenge, then, is no longer simply deciding where to host an application, but ensuring easy access, smooth data flows, and a coherent view of the entire infrastructure.
According to Gartner, spending on sovereign cloud infrastructures is expected to rise by 35.6% in 2026. This increase reflects a desire to better control data, governance, and technological dependencies.
In France, critical services such as the SAMU information system and IGN’s Geographical Portal are hosted on French clouds. Cloud infrastructures now underpin essential public services. The cloud, the network, and security can no longer be considered separately: from the user’s perspective, they form a single chain.
Multicloud alone does not guarantee service continuity
Working with multiple providers reduces certain dependency risks, but it does not by itself guarantee continuity of service. An application may be deployed on two clouds and still be exposed to a single point of failure if those clouds, sovereign or not, are not properly interconnected.
A secondary environment must therefore be available and ready to take over connectivity in case of failure. Continuity must be designed, maintained, and regularly tested. This involves data synchronization, access configuration, and testing failover mechanisms. Resilience must be considered end-to-end, taking into account the relationships between applications, data, users, clouds, and networks.
Companies are increasingly looking for a centralized orchestration layer that provides unified visibility, policy management, and automation and control capabilities across cloud and network environments. They must be able to detect performance degradations, quickly identify the cause, and re-route traffic with minimal manual intervention.
Real-time link monitoring with automatic rerouting enables switchover in seconds rather than after manual intervention. Without this level of end-to-end centralized orchestration, multicloud offers redundancy, but does not guarantee effective service continuity.
The network becomes the unifying layer
In a distributed architecture, the network connects users, sites, data centers, and clouds, while providing visibility into traffic flows. When an application slows down, organizations must determine whether the issue stems from the platform, the connection, a security policy, or the application itself.
A unified network layer enables a company to apply consistent measures for segmentation, encryption, and access control across different environments. Thus, a workload moving from one cloud to another does not end up with a reduced security level during the transfer.
The network must also adapt to different use cases. Collaboration tools, industrial applications, and AI platforms do not generate the same data volumes and do not have the same latency requirements. An intelligent networking approach can optimize traffic flows and cloud connectivity, prioritizing critical services and automatically rerouting workloads when performance deteriorates.
This improves application performance while reducing unnecessary data transfers and egress costs. SD-WAN, SASE, and Zero Trust approaches can help meet these needs, provided they are managed coherently. Deployed in silos, they add another layer of complexity.
From default multicloud to a multicloud designed with purpose
AI further heightens interdependencies. An AI project typically relies on data stored across multiple locations and computational resources provided by several vendors. Model performance alone is not enough: if data cannot move efficiently, processing delays and costs rise as the project scales.
Managing multiple clouds and hybrid environments requires specialized expertise, which drives up operating costs, lengthens problem-resolution times, and increases risk for organizations.
The next step for enterprises is therefore to make their existing cloud environments work better together through unified management. By combining orchestration, automation, and continuous optimization, organizations gain agility, use network resources more efficiently, and curb operational expenses.
Resilience does not depend on the number of platforms. It hinges on the ability to connect, observe, and orchestrate them as a single, coherent environment.
*Amit Kapoor is Vice President and Head of the Europe Division at Tata Communications
