Hybrid Multi-Cloud and AI Workloads Reshaping Enterprise IT Infrastructure: Analysis of HPE and Microsoft Executive Roundtable Discussion

Hybrid Multi-Cloud and AI Workloads Reshape Enterprise IT Infrastructure: Analysis of HPE and Microsoft Executive Roundtable Discussion

In June 2026, an iTnews executive roundtable luncheon sponsored jointly by HPE and Microsoft was held at Aria Restaurant in Sydney. IT leaders from Jefferies Australia, the University of Sydney, QBE, Toll Group, InfraBuild, Worldline, Challenger, HLB Mann Judd, Crown Resorts, and other organizations gathered to discuss the core theme: "How to modernize enterprise IT infrastructure to private, hybrid, and multi-cloud environments," with special emphasis on reducing dependency risks, controlling costs, improving flexibility, and addressing the surge in AI workloads. This is not only a concern for technology leaders in Australian enterprises but also a paradigm shift underway in the global cloud computing market.

Event Background: The Turning Point from "Migrating to the Cloud" to "Optimizing the Cloud"

Over the past decade, "migrating to the cloud" has been the main theme of digital transformation for enterprises. But today, the pure public cloud strategy is facing challenges—cost runaway, vendor lock-in, compliance complexity, and the special requirements of AI training and inference workloads are prompting companies to re-evaluate their architectures. The joint HPE and Microsoft roundtable exactly occurred at this turning point: the hybrid infrastructure solutions (such as HPE GreenLake and Microsoft Azure Arc/Azure Local) provided through their collaboration are becoming the template for enterprises to address new demands.

Technical Analysis: How Hybrid Multi-Cloud Architecture Supports AI Workloads

#### Hybrid Multi-Cloud: From "Optional" to "Mandatory"

Hybrid multi-cloud architecture is not a new concept, but AI workloads have pushed it to new heights. AI training requires massive GPU computing power, often provided on-demand by public clouds; while inference tasks may need to be closer to data sources, requiring on-premises deployment to avoid latency and data privacy risks. The as-a-service model offered by HPE GreenLake allows enterprises to run infrastructure on-premises with a consumption model similar to public clouds, while unified management via Azure Arc—this is a typical manifestation of the "edge-core-cloud" continuum in the AI era.

#### Azure Local: Redefining the Localized Cloud Experience

"Azure Local" (evolved from Azure Stack HCI), repeatedly mentioned in the roundtable, is a key technology from Microsoft. It enables enterprises to deploy cloud services consistent with Azure—including Kubernetes, databases, and AI inference services—at the edge or in data centers. Combined with HPE's hardware optimizations (such as AI-optimized ProLiant servers, storage, and networking), enterprises can build a seamless hybrid environment that allows AI workloads to run efficiently anywhere.

Core Issues Addressed: Vendor Lock-In and Cost Control

In the roundtable discussion, "reducing dependency risks" was listed as a top priority.#### Core Problem Solved: Vendor Lock-in and Cost Control

During the roundtable discussion, "reducing dependency risk" was listed as a priority. The hybrid multi-cloud strategy reduces lock-in through multi-cloud interconnection (e.g., Azure Arc supports AWS and GCP resource management), while HPE GreenLake's pay-as-you-go model avoids over-provisioning. According to Gartner, by 2028, 70% of enterprises will adopt hybrid multi-cloud management platforms, with the integration of AI workload optimization becoming a key driver.

Enterprise Impact Analysis: Cost, Operations, and AI Readiness

#### Cost Impact: Shift from CAPEX to OPEX, but Requires Fine Management

The as-a-service model of HPE GreenLake and Azure Local allows enterprises to shift capital expenditure (CAPEX) to operational expenditure (OPEX), alleviating cash flow pressure. However, roundtable participants pointed out that without resource planning for AI workloads, OPEX may exceed expectations. For example, the power and cooling costs of GPU clusters in private deployments may be higher than anticipated. Enterprises need to use FinOps tools to allocate costs to business units.

#### Deployment and Operations: The Value of a Unified Management Layer

The biggest operational challenge of hybrid multi-cloud is managing multiple heterogeneous environments. Azure Arc provides a unified control plane, while HPE GreenLake Central offers hardware-level monitoring. In the roundtable, Jonathan Woods of QBE emphasized the importance of automated operations—adopting Infrastructure as Code (IaC) and AIOps tools can reduce manual intervention.

#### Security and Compliance: Rising Demand for Sovereign Clouds

Countries like Australia have increasingly strict data sovereignty requirements. Azure Local supports data residency, and HPE also provides on-premises encryption key management. Jihad Zein of Toll Group pointed out that supply chain data in the logistics industry involves cross-border regulations, and a hybrid architecture can ensure compliance without sacrificing AI capabilities.

Market Competition Analysis: Who Will Benefit?

#### Cloud Vendor Competition: Transformation Pressure on AWS, Azure, Google Cloud

AWS Outposts and Google Distributed Cloud are similar, but the Azure Arc+Azure Local ecosystem is more complete, especially as the partnership with HPE fills the hardware layer capabilities. HPE, as a neutral hardware vendor, can support multi-cloud more flexibly, which is attractive to enterprises that reject single cloud lock-in.

#### Data Center and Infrastructure Providers: Equinix, Digital RealtyAI workloads are increasing demand for data center density and liquid cooling, benefiting colocation providers like Equinix. However, HPE's GreenLake solution may divert some of the colocation demand, as enterprises can deploy infrastructure on their own premises.

#### SaaS and AI Platforms: Impact on Oracle and SAP

Traditional SaaS vendors that fail to offer flexible hybrid deployment options will be eroded by "platform + hardware" combinations like HPE+Microsoft. For example, SQL Server or AI inference services running on Azure Local may replace some SaaS functions.

Industry Trend Observations: AI Native Cloud and Infrastructure as a Service

#### The Rise of AI Native Cloud

The most inspiring trend from the roundtable is "AI Native Cloud"—infrastructure optimized for AI from the ground up. For instance, HPE's Cray EX supercomputing technology is being integrated into GreenLake; Microsoft's Azure ND series GPU instances are also being localized. In the next five years, any cloud platform without built-in AI acceleration capabilities will lose competitiveness.

#### Green Data Centers and Energy Efficiency

The high energy consumption of AI training has drawn attention. Both HPE and Microsoft have committed to using renewable energy and adopting liquid cooling technology. Roundtable participants believe that energy efficiency will become a key KPI when enterprises choose infrastructure providers.

#### Edge AI and Industry Integration

The participation of industrial companies like InfraBuild indicates that AI inference is rapidly penetrating vertical industries such as manufacturing and logistics. Running pre-trained models on edge nodes requires compact yet powerful hardware—exactly where HPE's edge solutions and Azure IoT Edge come into play.

CloudTechDaily Insight

Although this roundtable was positioned as a small discussion, the topics reflect the profound transformation taking place in enterprise IT infrastructure: AI is no longer just a workload in the cloud but the core driver of architectural redesign. The collaboration model between HPE and Microsoft—hardware as a service plus cloud management plane—is likely to become the standard paradigm for the next decade. For enterprise CIOs, the key takeaway is that hybrid multi-cloud is not a compromise but a strategic choice to maximize flexibility, control, and cost efficiency in the AI era. Relying solely on public cloud or pure on-premises deployment is unsustainable; instead, a hybrid architecture based on a "consistent cloud experience" should be built. At the same time, enterprises must plan their AI workload distribution strategy in advance, layering training, inference, and data storage, and leveraging FinOps for granular cost management. In the future, infrastructure providers that can seamlessly integrate computing power, data, and AI capabilities will gain a significant first-mover advantage in the market.*来源：iTnews - "In Pictures: HPE & Microsoft IT infrastructure roundtable"*