For the past few years, the industry narrative suggested that the future of enterprise IT was a one-way street toward the public cloud. But across Asia Pacific, Japan (APJ), and the Middle East, it is clear we have moved past simple hybrid cloud adoption. Last year, our data tracked the "Cloud Reset", a deliberate, tactical rebalancing of workloads. In 2026, generative and agentic AI have accelerated that reset into The AI Tipping Point.
Our latest report, Private Cloud Outlook 2026 - The AI Tipping Point, captures insights from 1,800 IT leaders globally (including 600 in APJ), and confirms this definitive pivot. Enterprises are realizing that while the public cloud is ideal for early AI experimentation, private cloud is becoming the preferred platform when it comes time to scale production AI. The shift is being shaped by three forces — costs, complexity, and control — that public cloud environments are increasingly failing to address for production AI at scale. This AI inflection point is accelerating a broader trend: an overwhelming 82% of organizations in Asia Pacific and Japan are considering workload repatriation to private cloud, while 54% have already repatriated some workloads.
Public Cloud Waste Hits a Breaking Point
The motive behind this shift is economics. Globally and across Asia Pacific and Japan cost management is named as the number one public cloud challenge by more than 30% of survey respondents. The inefficiency has reached a global breaking point with 97% of all enterprise leaders admitting that a portion of their public cloud spend is wasted. This is even more pronounced in our region where 57% of organizations experience public cloud waste exceeding a quarter of their entire cloud budget; the highest globally.
APJ enterprises are realizing they cannot scale AI sustainably when a quarter or more of their cloud budget is disappearing into wasted resources. This public cloud friction is a primary driver behind the move to private cloud - with cost predictability (39%) and performance demands (39%) ranking as the top catalysts for workload repatriation globally.
The Sovereignty Imperative: Moving Past Broad Compliance
Operating in the Asia Pacific Japan and Middle East region means navigating a highly fragmented, rapidly evolving regulatory landscape. However, the conversation has fundamentally shifted. The report reveals that 87% of IT leaders in APJ state that localized regulatory shifts directly impact their IT strategy.
Whether navigating Australia’s Security of Critical Infrastructure Act, India’s Digital Personal Data Protection (DPDP) Act , or Japan’s Act on the Protection of Personal Information (APPI), governments are demanding that citizen data remain within physical borders. Keeping data within a private cloud isn't just a security preference anymore; it is an absolute requirement for maintaining operational continuity.
Localized Realities: Ransomware in Japan, Skills Shortages in India, and Operational Silos in Australia
When we zoomed into individual markets, we observed that in Japan, infrastructure strategies are being reshaped by defensive data protection parameters. The region has seen an aggressive spike in sophisticated cybersecurity threats, making the ransomware trend in Japan a board-level crisis. In addition, this change in the threat landscape reinforces our global finding that data protection and privacy (37%) along with security and control (36%) are the top new demands placed on IT by AI.
In India, the primary obstacle is a significant deficit in the specialized talent required for AI infrastructure and operations. With India’s large developer community, there is tremendous opportunity to invest in architectural expertise needed to manage fragmented, multi-cloud AI environments. This localized challenge also reflects a wider trend across the APJ region, where the talent gap is most pronounced in AI infrastructure operations (39%) and cloud-native Kubernetes management (38%). Consequently, 84% of APJ organizations are compelled to depend on outsourcing or professional services to manage the skills gap.
Within Australia, we are witnessing a highly sophisticated and mature cloud market leading the regional charge in bringing workloads back home. Early public cloud adopters in the Australian enterprises are encountering the commercial realities of AI inference workloads in public cloud ecosystems. However, their primary hurdle isn't technology, it is structure. Within Australian organisations, 35% of enterprise IT leaders cite skills gaps and 32% cite siloed IT teams as the greatest barriers to private cloud adoption. For these mature enterprises, the move to private infrastructure is less about physical server migration and more about a cultural shift toward unified management. They are deploying automated private clouds specifically to bridge internal gaps, ensuring strict alignment with national frameworks on critical asset protection and domestic data governance without inflating operational overheads.
To combat this widespread regional complexity, 73% of APJ organizations are evolving their platform engineering teams, breaking down isolated technology silos and consolidating compute, storage, and networking into a single, automated private platform that existing teams can actually manage.
The Modernization Journey: Upgrading the Cloud Operating Model
Building, optimising and repatriating AI inference successfully requires a comprehensive private cloud modernization journey. True modernization means shifting from fragmented infrastructure components to a unified, software-defined platform that delivers a cloud operating model on-premises.
Enterprises are shifting their strategic priorities to support this shift. According to the report, 58% of IT leaders now state that building new workloads directly on a modernized private cloud is a top priority, a steady climb from 53% just a year ago. Furthermore, organizations are decisively putting their money where their strategy is: over a three-year outlook, private cloud spending intent has risen by 21 percentage points, growing at more than double the rate of public cloud spending intent (which rose by just 10 points). This realignment proves that modernization isn't an experimental phase; it is an active investment in building robust, local private platforms optimized for virtual machine and containerized workloads.
For our customers, this data validates exactly why we built VMware Cloud Foundation (VCF) and evolved it for modern and AI workloads. Central to this evolution is driving breakthroughs in hardware efficiency with a unified private cloud platform with integrated compute, storage, networking, security, and management across all endpoints. According to Kotak Mahindra Bank, one of the leading banks in India, "Evaluating VCF 9.0 Memory Tiering demonstrated how extending DRAM with NVMe improves memory utilization and optimizes infrastructure investments."
The old trade-off to sacrifice public cloud agility for private cloud control is obsolete. Standardizing on a modernized platform like VCF gives developers the self-service, API-driven velocity of a public hyperscaler, while giving the business the cost predictability, data sovereignty, and security required to run AI safely.
The experimental phase of AI is over. To scale production AI without breaking the budget, suffering talent burnout, or violating localized data sovereignty laws, the logical step is to place workloads in the environment best suited to business requirements. Let’s talk about how we can design a modern private cloud architecture built for the long-term future of your enterprise.