Enterprise AI finds its infrastructure home in private cloud as costs and security reshape strategies

The AI tipping point: where enterprise AI runs at scale

When enterprises first began building AI strategies, the default assumption was straightforward: AI would run in the hyperscaler cloud. The APIs were ready, GPU capacity was building out, and the inertia of a decade of public cloud investment pointed in one direction. Broadcom's Private Cloud Outlook 2026 report finds that, as enterprises move to scale, the direction has changed. The Private Cloud Outlook 2026: The AI Tipping Point draws on a blind, global survey of 1,800 senior IT leaders across eight countries. Now in its second year, the report tracks a shift in cloud strategy that is no longer something on the horizon, but one already showing up in production workloads, capital budgets, and board-level priorities. Enterprise AI has found its infrastructure home in private cloud. Production AI is moving to private cloud Last year, 56 percent of enterprises used public cloud as the primary environment for production AI inference. This year, that figure has fallen 15 percentage points to 41 percent, while 56 percent of enterprises are now running or planning to run production inferencing in a private cloud. The shift goes deeper than the top-line numbers. Forty-three percent of enterprises actively repatriating workloads are moving AI training, large language models, and inference out of the public cloud, a category that did not exist in last year's study. The broader repatriation trend has accelerated sharply as well: 83 percent of enterprises are now considering repatriation , up from 69 percent in 2025, and half have already moved at least some workloads, a 15-point jump in a single year. The forces driving enterprise AI to private cloud are the same ones that pulled storage, security-sensitive applications, and regulated data there before it. Security, control, cost, and governance did not become more important because of AI, but the consequences of getting them wrong became much harder to absorb at production scale. When IT leaders place workloads, those classified as high-security, latency sensitive, business critical, or data-intensive consistently land in private cloud. The bill for AI infrastructure has arrived For the first time in this study, cost has overtaken security as the top concern about public cloud. That reflects a familiar reality for enterprise IT leaders: public cloud costs were already difficult to forecast and manage, and AI workloads have made that problem substantially worse. Nearly all IT leaders surveyed (97 percent) believe some portion of their public cloud spend is wasted, and more than half (52 percent) say that waste exceeds 25 percent of their total spending. Generative AI and agentic workloads are compounding the pressure, with 62 percent of IT leaders reporting that they are very or extremely concerned about AI infrastructure costs. Enterprises are revising their investment strategies accordingly. Net intent to increase private cloud investment over three years has risen from 51 percent to 72 percent, and private cloud investment is now growing at more than twice the rate of public cloud. Cost predictability has become the second biggest driver of that shift, cited by 39 percent of organizations. Enterprises that built AI ambitions on variable, consumption-based public cloud pricing are recalculating. Private cloud, with its predictable economics and direct IT control over infrastructure, is increasingly where the budget decisions are landing. Sovereignty has become a board-level priority Geopolitics has moved squarely into the infrastructure conversation. Eighty-six percent of IT leaders say geopolitical and regulatory factors are now directly affecting their IT strategy and operations. Data sovereignty and residency requirements are the top concern, cited by 54 percent of respondents, followed by jurisdiction-specific compliance requirements at 51 percent. For enterprises operating across borders, decisions about where data lives carry direct implications for where workloads can run. AI workloads that process sensitive, regulated, or proprietary data require infrastructure that provides governance and control from the ground up. Security and compliance remain the single most important factor in workload placement decisions, cited by 32 percent of respondents. AI is adding new obligations on top of existing ones: data protection and privacy (37 percent) and security and control (36 percent) are now the leading infrastructure requirements that AI imposes. Private cloud provides the governance architecture to meet those requirements by design, built in from the start rather than bolted on after deployment. Complexity is a platform problem Running production AI at enterprise scale is an operations challenge as much as an infrastructure one. The top skills gap cited by IT leaders is AI infrastructure and operations, named by 40 percent of respondents, followed by cloud security operations at 38 percent and Kubernetes operations at 37 percent. To close that gap, 81 percent of enterprises now fully outsource or use professional services for their cloud-related needs. Operational simplification matters as much as picking the right technology partners. Enterprises that standardize on a unified, well-governed private cloud platform address the AI skills challenge with fewer specialists, less operational fragmentation, and clearer organizational accountability. A platform-centric approach reduces the surface area that teams have to manage, and that is where the real operational gains lie. The tipping point is here The Private Cloud Outlook 2026 confirms what the data has been building toward for two years. Enterprise IT has reached the AI tipping point, and private cloud is the preferred platform for production AI because it addresses what AI at scale demands: security, cost predictability, data sovereignty, and governance that enterprises cannot treat as optional. VMware Cloud Foundation 9.1 is built for this environment. It provides a unified platform for running AI and traditional workloads together, with the performance, cost controls, and security capabilities that production AI at enterprise scale requires. The research shows where enterprise AI is heading, and VMware Cloud Foundation is the platform built to get organizations there. Read the full Private Cloud Outlook 2026 report: https://www.vmware.com/docs/private-cloud-outlook-2026

Private cloud strengthens production AI security

Three insights you might have missed from theCUBE's coverage of Broadcom's 'Modern Private Cloud' event Enterprises are rapidly pushing AI projects into production in the race for higher revenues - but it's not without cost and security risks. Private cloud is emerging as a bedrock for AI infrastructure as businesses seek control, cost predictability, security, and compliance. Broadcom Inc. is positioning its VMware Cloud Foundation platform as the secure, cost-effective private cloud foundation enterprises need to effectively seize new AI opportunities, explained Paul Turner (pictured), chief product officer of the VMware Cloud Foundation Division at Broadcom Inc, in a recent interview with theCUBE's John Furrier and Gemma Allen. "AI is also a risk and cost multiplier," he said. "Just think about a few stats: 73% of enterprises see AI-related attacks. That is almost every industry out there ... actually seeing these new attacks that are driven by AI-enabled software." Turner was among several industry experts from Broadcom, ThinkOn and Charlotte Pipe and Foundry Co. who spoke to Furrier and Allen at the Broadcom "Modern Private Cloud: A Secure Foundation for Production AI" event, during an exclusive broadcast on theCUBE, SiliconANGLE Media's livestreaming studio. They discussed.. (* Disclosure below.) Here are three insights you may have missed from theCUBE's coverage of Broadcom's "Modern Private Cloud: A Secure Foundation for Production AI" event: Insight #1: Rising AI infrastructure costs fuel the private cloud shift AI offers enticing revenue potential for enterprises, but those benefits are cut across by high infrastructure costs. The search for cost efficiency is driving rapid adoption of the VCF platform, explained Turner, with more than 2,000 customers already on board. "I mean, it is getting very, very expensive to start running services," he told theCUBE. "Companies are suddenly seeing the cloud answer to AI is not the right answer to run production AI at scale, and inferencing AI, which is really the runtime of your day-to-day operations. You will not run that on the cloud, because of the cost of operations. You can run it more efficiently, just like you can the rest of your infrastructure, on a private cloud environment." Tokenomics is prompting executives to reconsider where workloads should run. Security concerns are also influencing whether they choose the public cloud or private cloud, added Prashanth Shenoy, chief marketing officer and vice president of marketing of the VMware Cloud Foundation Division at Broadcom "Last year, when we did the private cloud outlook study, there was a definitive cloud reset happening in the market, where private cloud and the operating model of private cloud to run your mission-critical workload on-premises or in a hybrid environment was on par with public cloud," he said. "Fast-forward to this year, when we did the same survey with 1,800 IT leaders and decision-makers around the globe. A lot of organizations are now moving their AI applications from a pilot phase of trying, to production, doing it at scale." Here's theCUBE's complete video interview with Paul Turner and Prashanth Shenoy: Insight #2: AI sovereignty now encompasses infrastructure, data and models Sovereignty is increasingly a priority as organizations seek greater freedom and control over how they run AI. Many also want to reduce their dependence on public cloud providers. The scope of AI sovereignty in the enterprise has shifted in just the past year, explained Chris Wolf, global head of AI and advanced services for the VMware Cloud Foundation Division at Broadcom. Whereas a year ago organizations may have primarily been focused on filtering the data they shared with frontier models, those same businesses today are making even more nuanced and mindful decisions around data privacy, access control, and auditability. "For a lot of our customers today, their definition means that it's not just about the data plane being sovereign, it's about the control plane being sovereign," he said in an interview with theCUBE. "It's that I can disconnect from the internet and I can continue to run. I can continue to operate. That's a difference and that's been really driven over the last couple of years, far more so than we've seen previously." Sovereignty is especially a priority for government customers. ThinkOn, a Canadian cloud services provider, worked with Broadcom to launch Canada's first sovereign cloud. In this case, private cloud was the only viable infrastructure option able to deliver the accountability, auditability and compliance required in the government sector, explained Craig McLellan, founder and chief executive officer of ThinkOn. Additionally, the need for sovereignty extends beyond the cloud down to the AI models. "I'd even go a step further and say that it's also about model sovereignty," he added. "Many countries want to have their own sovereign models. For instance, in Canada, Cohere is a vibrant participant in the market, and we actually took the opportunity to work closely with Broadcom to add the Cohere model to the environment as a private cloud. We are able to provide the public sector with a combination of model sovereignty, certainly economic and data sovereignty, as well as control plan and data plane sovereignty." Here's theCUBE's complete video interview with Craig McLellan and Chris Wolf: Insight #3: Enterprises that modernized first are best positioned for AI integration Businesses that have already modernized their infrastructure in the private cloud may be in a better position to adopt AI in the future. For example, Charlotte Pipe and Foundry, a 125-year-old PVC and cast iron pipes and fittings manufacturer, invested in Broadcom's VCF years ago as a way to optimize workloads and security, noted Rodney Barnhardt, server administration at Charlotte Pipe and Foundry. "Originally when we moved to VCF, it was prior to being able to do brownfield imports," he told theCUBE. "While we've been VMware customers for a long time, prior to moving to VCF, we were on three tiers: Cisco, BladeCenter, Unity all-flash storage array. In looking at VMware by Broadcom and the VCF platform using HCX to be able to do those migrations, as well as vDefend to put microsegmentation around products, made VMware Cloud Foundation an ideal product to look at deploying within our environment." As AI adoption expands across the enterprise, greater oversight is needed to manage the increasing number of workflows interacting with sensitive operational data. Security capabilities like vDefend, which enables microsegmentation, help reduce the ability for external threats to move laterally throughout Charlotte Pipe's IT environment, Barnhardt pointed out. That's especially important because connectivity, and the ability for internal systems to talk to one another, could be the ultimate benefit of a modern AI-enabled infrastructure. "As more organizations look to AI, they'll be connecting various systems that may now be disconnected," Barnhardt said. "They'll be looking at integrating those better and allowing them to rely on each other and say, 'This system may take an order in and send it to another system that processes the order that may then generate something on the processing side or deployment side.' So I think you'll see more integrations like that and discussions between teams on doing those types of integrations. Here's theCUBE's complete video interview with Rodney Barnhardt: Catch up on our complete video coverage of the Broadcom "Modern Private Cloud: A Secure Foundation for Production AI" event (* Disclosure: TheCUBE is a paid media partner for the Broadcom "Modern Private Cloud: A Secure Foundation for Production AI" event . Sponsors of theCUBE's event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)