Production Inference Shifts Decisively to Private Cloud

The AI experimentation phase is over, and the private cloud is where enterprise AI workloads are being deployed for security and scale, according to Private Cloud Outlook 2026, a new report from Broadcom.

Where last year's report documented a deliberate "cloud reset" toward balance between public and private cloud, 2026 marks an acceleration into a full AI tipping point. 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.

Key findings from the report include:

  • 56% of enterprises are running or planning to run production AI inferencing on private cloud, while public cloud use for the same workloads dropped 15 percentage points year over year — from 56% to 41%.
  • The biggest new demands being placed on enterprise IT by AI are data protection and privacy (37%) and security and control (36%).
  • For the first time, cost has overtaken security as the number one public cloud concern, rising from 26% in 2025 to 31% in 2026.
  • 97% of IT leaders believe some of their public cloud spend is wasted, and 52% estimate that waste exceeds 25% of their total public cloud budget.
  • 83% of enterprises are considering the repatriation of workloads from public to private cloud, and 50% have already done so — with cost predictability now jumping to second biggest driver for repatriation, cited by 39% of organizations.
  • Four out of five IT leaders say geopolitics are now affecting their IT strategy and operations, and for the first time, data sovereignty and residency requirements (54%) have overtaken jurisdiction-specific compliance (51%) as the leading geopolitical factor shaping infrastructure decisions.

"As enterprises move from pilots to running AI at production scale, infrastructure and operational costs spike, security gaps surface, and complexity compounds. The research is clear: enterprises increasingly prefer private cloud for production AI," said Prashanth Shenoy, VP of Marketing, VMware Cloud Foundation Division at Broadcom.

The Inference Shift: Production AI Finds a Home on Private Cloud

The defining data point of this year's report is the scale and speed of the shift in where enterprises are running AI workloads. While public cloud remains a viable environment for AI pilots and model training experiments, the economics of running inference at scale tell a different story. 56% of enterprises are running or planning to run production inferencing on private cloud, compared to just 41% on public cloud — a reversal from last year's near-parity. The drop of 15 percentage points in public cloud's share of production AI workloads in a single year is among the most dramatic shifts in this year's report.

The reason is straightforward. As IT leaders described it in the survey data: public cloud is too expensive and insufficiently governed for AI workloads at scale. For pilots and training, some agility trade-offs may be acceptable. But when organizations need to scale, the cost and governance requirements drive workloads back home. 62% of IT leaders report being very or extremely concerned about the infrastructure costs of running generative and agentic AI, while 36% say AI is actively driving new requirements for data protection, privacy, security controls, and risk management.

The Sovereignty Mandate: Geopolitics Reshapes Infrastructure Strategy

Geopolitics has entered the infrastructure conversation in 2026. Four out of five IT leaders now report geopolitics are directly affecting their IT strategy and operations. Data sovereignty has moved from a compliance checkbox to a board-level priority, and data sovereignty and residency requirements (54%) have overtaken jurisdiction-specific compliance (51%) as the leading geopolitical factor shaping infrastructure decisions. Industries with high security and compliance requirements — financial services, public sector, healthcare, and life sciences — are at the leading edge of this shift. For these organizations, the combination of AI-driven data volumes, cross-border data governance complexity, and the rising cost and governance burden of public cloud is making a compelling case for private cloud infrastructure that helps keep sensitive data under organizational control.

The Cost Reckoning: Public Cloud Economics Are Breaking Down

Cost has now become the defining concern about public cloud, overtaking security as the number one public cloud challenge, jumping from 26% in 2025 to 31% in 2026. The waste figures are striking: 97% of IT leaders believe some portion of their public cloud spend is wasted, and 52% believe that waste exceeds 25% of their total public cloud budget.

These economics are directly translating into repatriation activity. 83% of enterprises are now considering moving workloads from public to private cloud, and 50% have already repatriated at least some workloads. The top three drivers of repatriation are security and compliance (51%), cost predictability (39%) and performance (39%). The explosive rise of cost predictability as a repatriation driver is one of the most significant year-over-year changes in the report, underscoring how severely public cloud economics have deteriorated in the AI era.

Against this backdrop, private cloud investment intent is accelerating. Private cloud spend intent is growing at twice the rate of public cloud — up 21 points versus 10 points over a three-year outlook — and 58% of IT leaders now name building new workloads on private cloud as a top priority, up from 53% one year ago.

Methodology: The Private Cloud Outlook 2026 is based on a global survey conducted by Radius Tech in partnership with Broadcom. The survey was fielded in February–March 2026 and included 1,800 senior IT decision-makers at enterprise organizations (1,000 or more employees) across eight countries in North America, Europe, and Asia-Pacific. The report was published in June 2026.

Related Links

Hot Topics

The Latest

When the Database Is the Problem, Is APM Enough?

Production incidents rarely announce themselves as database problems. They appear as slow transactions, timeouts, rising response times, or an application struggling under a workload it previously handled. APM provides an essential starting point. It can identify a slow transaction path, highlight an affected service, and show that a database dependency is consuming more time than expected. But identifying the database as part of the problem is not the same as explaining what is happening inside it ...

From Cleanup to Prevention: The Future of Cloud-Native Cost Optimization

Cloud teams are under constant pressure to reduce spend without slowing development or increasing operational risk. They are deploying autoscalers, rightsizing workloads, enforcing resource requests, reviewing utilization dashboards, and building FinOps processes around cloud-native environments. Yet the results often disappoint ...

The Real Enterprise AI Risk Isn't Hallucinations

Ask most IT leaders about their biggest concern with AI and you'll hear the same answer: hallucinations ... Today, however, the conversation has shifted ... As organizations move beyond chatbots and experiments, they are increasingly deploying AI agents that perform multi-step tasks. These systems retrieve documents, query databases, call APIs, generate reports, write code, and make recommendations. The issue is not whether the model can reason. The issue is whether the organization can see, verify, and govern the decisions being made along the way ...

Beyond Buy vs. Build: The New Playbook for Data Pipelines

While organizations want to take control of their telemetry, building telemetry pipelines from scratch can be a very daunting, complicated task, even when leveraging open-source standards like OpenTelemetry. It requires specialized knowledge across distributed systems, data engineering, and security. This fragmented approach across systems causes higher operational costs; it puts a strain on resources and reduces efficiency as teams have to work with different interfaces and processes ...

The Corporate Network Is Fading - Here's What Replaces It

For decades, enterprise networks were designed around a simple assumption: work happened inside the office. Applications lived in centralized data centers, employees connected through internal infrastructure, and security focused on protecting the perimeter that surrounded everything ... But the way organizations operate today bears little resemblance to that environment. Cloud platforms host critical applications, employees connect from homes and airports as often as they do from offices, and partners collaborate through shared systems that exist far beyond corporate walls. In short, the corporate network no longer resembles the environment it was designed to protect ...

IT Leaders Must Establish a Network Data Architecture Practice

As an analyst who researches how IT organizations design, build, and operate their networks, I find that network data is a constant source of pain. Network teams struggle with data quality, fragmentation, authority, access, and trust. And these issues undermine everything they try to do. Here are the numbers: Only 45% of network teams are completely confident in the accuracy of their network source of truth, which documents the intent of their network ...

Data Center Outages Are Decreasing but More Costly

The 2026 Global Data Center Survey from Uptime Institute reveals an industry navigating workforce constraints, escalating outage expenses, even as rising costs remain the top concern for management teams ...

Observability Has a Physical Layer Now: What AI Infrastructure Teaches IT Operations

The next observability gap may not be in the code. It may be under the rack. That sounds strange until you think about how AI incidents actually feel in the middle of an investigation ... The application dashboard may be accurate. It may also be stopping at the wrong boundary. AI systems depend on software, but they also depend on a dense physical stack: racks, power paths, thermal margin, maintenance activity and, in many environments, liquid cooling. Those physical dependencies can change slowly before they look like a software incident ...

TLS Certificate Expiration Is Becoming an Observability Problem

Certificate expiration is the rare outage you can see coming. Every TLS certificate carries the date it stops working, so the moment it will begin breaking connections is knowable in advance. That's what makes an expired certificate such a frustrating way to lose a service. What's changing now is how often that date comes around ...

Enterprise Cloud Observability: Choosing the Right Starting Point

Enterprises operate different combinations of workloads across cloud, hybrid and multicloud environments. For business-critical workloads, teams need to consider monitoring and observability early so they can detect health issues, investigate failures, and understand operational impact. Organizations place workloads on cloud platforms based on a combination of technical requirements, economics, existing dependencies, organizational standards, and business priorities. Their monitoring priorities therefore depend on what they operate and where those systems run. Those priorities will not look the same for every organization ...

Production Inference Shifts Decisively to Private Cloud

The AI experimentation phase is over, and the private cloud is where enterprise AI workloads are being deployed for security and scale, according to Private Cloud Outlook 2026, a new report from Broadcom.

Where last year's report documented a deliberate "cloud reset" toward balance between public and private cloud, 2026 marks an acceleration into a full AI tipping point. 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.

Key findings from the report include:

  • 56% of enterprises are running or planning to run production AI inferencing on private cloud, while public cloud use for the same workloads dropped 15 percentage points year over year — from 56% to 41%.
  • The biggest new demands being placed on enterprise IT by AI are data protection and privacy (37%) and security and control (36%).
  • For the first time, cost has overtaken security as the number one public cloud concern, rising from 26% in 2025 to 31% in 2026.
  • 97% of IT leaders believe some of their public cloud spend is wasted, and 52% estimate that waste exceeds 25% of their total public cloud budget.
  • 83% of enterprises are considering the repatriation of workloads from public to private cloud, and 50% have already done so — with cost predictability now jumping to second biggest driver for repatriation, cited by 39% of organizations.
  • Four out of five IT leaders say geopolitics are now affecting their IT strategy and operations, and for the first time, data sovereignty and residency requirements (54%) have overtaken jurisdiction-specific compliance (51%) as the leading geopolitical factor shaping infrastructure decisions.

"As enterprises move from pilots to running AI at production scale, infrastructure and operational costs spike, security gaps surface, and complexity compounds. The research is clear: enterprises increasingly prefer private cloud for production AI," said Prashanth Shenoy, VP of Marketing, VMware Cloud Foundation Division at Broadcom.

The Inference Shift: Production AI Finds a Home on Private Cloud

The defining data point of this year's report is the scale and speed of the shift in where enterprises are running AI workloads. While public cloud remains a viable environment for AI pilots and model training experiments, the economics of running inference at scale tell a different story. 56% of enterprises are running or planning to run production inferencing on private cloud, compared to just 41% on public cloud — a reversal from last year's near-parity. The drop of 15 percentage points in public cloud's share of production AI workloads in a single year is among the most dramatic shifts in this year's report.

The reason is straightforward. As IT leaders described it in the survey data: public cloud is too expensive and insufficiently governed for AI workloads at scale. For pilots and training, some agility trade-offs may be acceptable. But when organizations need to scale, the cost and governance requirements drive workloads back home. 62% of IT leaders report being very or extremely concerned about the infrastructure costs of running generative and agentic AI, while 36% say AI is actively driving new requirements for data protection, privacy, security controls, and risk management.

The Sovereignty Mandate: Geopolitics Reshapes Infrastructure Strategy

Geopolitics has entered the infrastructure conversation in 2026. Four out of five IT leaders now report geopolitics are directly affecting their IT strategy and operations. Data sovereignty has moved from a compliance checkbox to a board-level priority, and data sovereignty and residency requirements (54%) have overtaken jurisdiction-specific compliance (51%) as the leading geopolitical factor shaping infrastructure decisions. Industries with high security and compliance requirements — financial services, public sector, healthcare, and life sciences — are at the leading edge of this shift. For these organizations, the combination of AI-driven data volumes, cross-border data governance complexity, and the rising cost and governance burden of public cloud is making a compelling case for private cloud infrastructure that helps keep sensitive data under organizational control.

The Cost Reckoning: Public Cloud Economics Are Breaking Down

Cost has now become the defining concern about public cloud, overtaking security as the number one public cloud challenge, jumping from 26% in 2025 to 31% in 2026. The waste figures are striking: 97% of IT leaders believe some portion of their public cloud spend is wasted, and 52% believe that waste exceeds 25% of their total public cloud budget.

These economics are directly translating into repatriation activity. 83% of enterprises are now considering moving workloads from public to private cloud, and 50% have already repatriated at least some workloads. The top three drivers of repatriation are security and compliance (51%), cost predictability (39%) and performance (39%). The explosive rise of cost predictability as a repatriation driver is one of the most significant year-over-year changes in the report, underscoring how severely public cloud economics have deteriorated in the AI era.

Against this backdrop, private cloud investment intent is accelerating. Private cloud spend intent is growing at twice the rate of public cloud — up 21 points versus 10 points over a three-year outlook — and 58% of IT leaders now name building new workloads on private cloud as a top priority, up from 53% one year ago.

Methodology: The Private Cloud Outlook 2026 is based on a global survey conducted by Radius Tech in partnership with Broadcom. The survey was fielded in February–March 2026 and included 1,800 senior IT decision-makers at enterprise organizations (1,000 or more employees) across eight countries in North America, Europe, and Asia-Pacific. The report was published in June 2026.

Related Links

Hot Topics

The Latest

When the Database Is the Problem, Is APM Enough?

Production incidents rarely announce themselves as database problems. They appear as slow transactions, timeouts, rising response times, or an application struggling under a workload it previously handled. APM provides an essential starting point. It can identify a slow transaction path, highlight an affected service, and show that a database dependency is consuming more time than expected. But identifying the database as part of the problem is not the same as explaining what is happening inside it ...

From Cleanup to Prevention: The Future of Cloud-Native Cost Optimization

Cloud teams are under constant pressure to reduce spend without slowing development or increasing operational risk. They are deploying autoscalers, rightsizing workloads, enforcing resource requests, reviewing utilization dashboards, and building FinOps processes around cloud-native environments. Yet the results often disappoint ...

The Real Enterprise AI Risk Isn't Hallucinations

Ask most IT leaders about their biggest concern with AI and you'll hear the same answer: hallucinations ... Today, however, the conversation has shifted ... As organizations move beyond chatbots and experiments, they are increasingly deploying AI agents that perform multi-step tasks. These systems retrieve documents, query databases, call APIs, generate reports, write code, and make recommendations. The issue is not whether the model can reason. The issue is whether the organization can see, verify, and govern the decisions being made along the way ...

Beyond Buy vs. Build: The New Playbook for Data Pipelines

While organizations want to take control of their telemetry, building telemetry pipelines from scratch can be a very daunting, complicated task, even when leveraging open-source standards like OpenTelemetry. It requires specialized knowledge across distributed systems, data engineering, and security. This fragmented approach across systems causes higher operational costs; it puts a strain on resources and reduces efficiency as teams have to work with different interfaces and processes ...

The Corporate Network Is Fading - Here's What Replaces It

For decades, enterprise networks were designed around a simple assumption: work happened inside the office. Applications lived in centralized data centers, employees connected through internal infrastructure, and security focused on protecting the perimeter that surrounded everything ... But the way organizations operate today bears little resemblance to that environment. Cloud platforms host critical applications, employees connect from homes and airports as often as they do from offices, and partners collaborate through shared systems that exist far beyond corporate walls. In short, the corporate network no longer resembles the environment it was designed to protect ...

IT Leaders Must Establish a Network Data Architecture Practice

As an analyst who researches how IT organizations design, build, and operate their networks, I find that network data is a constant source of pain. Network teams struggle with data quality, fragmentation, authority, access, and trust. And these issues undermine everything they try to do. Here are the numbers: Only 45% of network teams are completely confident in the accuracy of their network source of truth, which documents the intent of their network ...

Data Center Outages Are Decreasing but More Costly

The 2026 Global Data Center Survey from Uptime Institute reveals an industry navigating workforce constraints, escalating outage expenses, even as rising costs remain the top concern for management teams ...

Observability Has a Physical Layer Now: What AI Infrastructure Teaches IT Operations

The next observability gap may not be in the code. It may be under the rack. That sounds strange until you think about how AI incidents actually feel in the middle of an investigation ... The application dashboard may be accurate. It may also be stopping at the wrong boundary. AI systems depend on software, but they also depend on a dense physical stack: racks, power paths, thermal margin, maintenance activity and, in many environments, liquid cooling. Those physical dependencies can change slowly before they look like a software incident ...

TLS Certificate Expiration Is Becoming an Observability Problem

Certificate expiration is the rare outage you can see coming. Every TLS certificate carries the date it stops working, so the moment it will begin breaking connections is knowable in advance. That's what makes an expired certificate such a frustrating way to lose a service. What's changing now is how often that date comes around ...

Enterprise Cloud Observability: Choosing the Right Starting Point

Enterprises operate different combinations of workloads across cloud, hybrid and multicloud environments. For business-critical workloads, teams need to consider monitoring and observability early so they can detect health issues, investigate failures, and understand operational impact. Organizations place workloads on cloud platforms based on a combination of technical requirements, economics, existing dependencies, organizational standards, and business priorities. Their monitoring priorities therefore depend on what they operate and where those systems run. Those priorities will not look the same for every organization ...