F5 Report Reveals 78% of Enterprises Now Run AI Inference as Core Operation
Because the Question Now Is Not Whether Companies Will Use AI, But Whether They Can Run It Reliably, Securely, and at Scale
F5, the global leader in delivering and securing applications and APIs, has released its annual State of Application Strategy (SOAS) Report, revealing that Artificial Intelligence (AI) has crossed a critical threshold: it is no longer an experimental initiative but a production workload demanding the same operational rigour as any mission‑critical system.
Based on responses from hundreds of enterprise IT and security leaders worldwide, the F5 research shows that 78% of organisations are now running AI inference themselves—a clear signal that enterprises are choosing control over convenience as AI becomes central to business operations. In APCJ, this figure stands at 50%.
Complexity at a New Inflection Point
The F5 findings arrive at a pivotal moment. With 93% of organisations operating across multiple clouds and 86% distributing applications across hybrid multicloud environments, the complexity of delivering and securing AI workloads has reached unprecedented levels.
“AI has moved from experimentation to operations. The question now is not whether companies will use AI, but whether they can run it reliably, securely, and at scale,” said Kunal Anand, Chief Product Officer at F5. “This year’s data shows a clear shift: AI inference is becoming core to the business, which means AI delivery is now a traffic management challenge, and AI security is now a governance and control challenge. The companies that understand this shift early will be the ones that move faster and more safely.”
Key Findings from the 2026 F5 Report
AI Is an Operational Reality
AI is no longer a flashy experiment or future concern; it has become an operational reality deeply embedded in daily business outcomes. Organisations now coordinate an average of seven AI models in production, with 77% reporting that inference—running trained models to generate outputs—has become their dominant AI activity, surpassing model building and training.
In APCJ, organisations use an average of three to four AI models in production, with 65% leveraging AI for real‑time operational automation. This shift underscores the need for operational governance, treating inference as a managed, policy‑driven workload integrated into the application stack and subject to the same architectural, security, and scalability demands as other production systems.
AI‑as‑a‑Service Strategies Are Risky
According to F5, AI‑as‑a‑Service strategies are increasingly viewed as risky and misaligned with modern enterprise realities. Only 8% of organisations globally rely exclusively on public AI services, F5 found. The majority are adopting hybrid deployment strategies across cloud, private, and self‑managed environments, building diversified model portfolios that require sophisticated routing, fallback, and policy controls to manage cost, accuracy, and availability.
Hybrid Multicloud Is the New Standard
Multicloud, multi‑environment operations are now the norm, with 93% of enterprises leveraging multicloud setups and 86% running applications across on‑premises, public cloud, and colocation environments. In APCJ, 87% of organisations use multiple cloud providers, 91% operate across multiple on‑premises data centres, and 87% leverage multiple colocation environments.
AI workloads mirror this complexity, requiring advanced routing, fallback, and policy controls to optimise cost, accuracy, and availability. A unified delivery, security, and governance strategy across environments is now essential to reduce silos, minimise operational disruptions, and maintain governance at scale.
AI Security and Governance Are Systemic Requirements
As AI systems enter full‑scale production, security has become an enterprise‑wide priority. The report shows 88% of organisations have faced AI‑related security challenges, while 98% are preparing for agentic AI—autonomous systems requiring identities, permissions, and guardrails akin to human users.
In APCJ, 54% of organisations cite the high cost of AI workloads as the top challenge, while 51% anticipate difficulties from the explosive growth of agent identities linked to agentic AI. This shifts the security perimeter to prompt, token, and identity layers, making governance across every layer essential.
Prompt and Token Layers as Control Points
The report highlights a significant shift in AI workload management, with control moving to prompts, tokens, and APIs. Nearly 29% of organisations identify prompt layers as the top delivery mechanism, while 23% prioritise token layers for delivery and security. In APCJ, 33% identify the prompt layer as the top mechanism, while around 23% prioritise the token/control layer. Governing these layers is key to optimising cost, performance, and safety.
Why It Matters
The 2026 State of Application Strategy Report by F5 offers a data‑driven view of the forces reshaping enterprise technology: the rapid operationalisation of AI, the permanence of hybrid multicloud, and an evolving threat landscape that demands new thinking about security and control.
AI maturity is quickly becoming a measurable indicator of operational resilience and competitive positioning. Organisations that invest in observability, authentication, and unified control across every environment where AI runs will be the ones that turn AI’s promise into lasting business value.
Download the full 2026 State of Application Strategy Report by F5 to access complete findings, industry benchmarks, and strategic recommendations.
