Google named Gartner cloud leader for ninth straight year

Google has been named a Leader in Gartner's 2026 Magic Quadrant for Strategic Cloud Platform Services, marking the ninth consecutive year it has received that designation.

Gartner placed Google furthest for Completeness of Vision in its assessment of strategic cloud platform providers. The result gives Google a fresh endorsement as competition in cloud infrastructure and artificial intelligence intensifies among the largest technology groups.

Google used the announcement to outline the infrastructure themes shaping customer demand, with a strong focus on AI systems, hybrid and multicloud operations, and tighter control over where data is stored and processed.

At the centre of that pitch is what Google describes as a co-designed technology stack spanning custom silicon, hardware systems, software, orchestration tools, models and agent-based applications. According to Google, its infrastructure teams work with Google DeepMind researchers on designs across hardware and software layers, including Tensor Processing Units, Arm-based Axion processors, Google Kubernetes Engine and Gemini models.

Google also sought to avoid the perception of a closed system. It pointed to work on open source and open standards, including llm-d for distributed inference, GKE Prism for benchmarking open models and TorchTPU for PyTorch compatibility across TPUs and GPUs.

That combination of components underpins Google's AI Hypercomputer offering. Google added that the same infrastructure foundation is used by nine of the top 10 AI labs globally.

AI demand

The update comes as cloud providers race to show they can support fast-growing AI workloads without forcing customers to overbuild expensive computing estates. Google said demand inside its own operations has risen sharply, with its data centres now processing 3.2 quadrillion tokens a month, about seven times more than a year earlier.

That increase helps explain the company's focus on what it calls dynamic infrastructure. The aim is to let customers shift between different types of compute, reserve capacity in advance for planned demand spikes and automate resource allocation as workloads change.

Google highlighted a mix of predefined and custom CPU shapes, Nvidia graphics processors, and its own TPUs and Axion CPUs. It also pointed to features in Google Kubernetes Engine, including Dynamic Workload Scheduler and dynamic resource allocation tools, as ways to improve utilisation and manage costs.

Another area of emphasis is application modernisation. Google said customers can use Gemini-based tools to map and update core applications, while Gemini Cloud Assist is designed to help identify and resolve operational issues in newer agent-based workflows.

Google is also trying to draw a link between AI adoption and network design. It said businesses increasingly want to connect enterprise systems and AI environments through a single control plane while retaining the flexibility to run workloads across hybrid and multicloud settings.

To support that argument, Google cited its Cross-Cloud Interconnect service and its private fibre backbone of more than 10 million kilometres. According to Google, testing showed network latency was more than 40% lower when traffic travelled over its cross-cloud network than when traffic to the same target was routed through the public internet.

Sovereignty focus

Google also devoted significant attention to digital sovereignty, an issue that has become more prominent for customers in regulated sectors and European markets. It argued that organisations should not have to choose between modernisation, advanced AI tools and regulatory control.

Its sovereign cloud approach is presented in three parts. The first is data sovereignty and control, using Google Cloud Data Boundary, External Key Management and Key Access Justifications to let customers define where data is processed and stored, and manage encryption keys outside Google infrastructure.

The second is local compliance and regional operations. Google said applications can run on physically and logically separated regional clouds operated by local partners for European customers. In France, S3NS provides PREMI3NS, a standalone sovereign cloud that has secured the SecNumCloud 3.2 qualification from the French National Agency for the Security of Information Systems.

The third strand is on-premises and air-gapped deployment through Google Distributed Cloud. Google said this gives customers a choice between a fully managed environment with no connection to the public internet and a connected model that keeps workloads on local hardware while linking to Google Cloud's central control plane.

No executive quotes accompanied the announcement, but the message was clear: Google wants to frame its cloud position around AI infrastructure, operational flexibility and customer control over data. Those themes are likely to remain central as cloud providers compete for large corporate workloads that are becoming more complex, more distributed, and more sensitive to cost and regulation.