Cloudflare will use LEAP 2026 in Riyadh to showcase an integrated approach to securing, governing and accelerating artificial intelligence workloads as organisations in Saudi Arabia move from AI experimentation toward production deployment.
The company plans to demonstrate technologies spanning Zero Trust security, edge computing, AI traffic visibility, large language model protection, content controls and infrastructure for autonomous AI agents.
Cloudflare said the focus reflects a growing shift in enterprise requirements as AI applications become more deeply embedded in business operations and introduce new risks around identity, data leakage, cost control and machine-to-machine activity.
AI security moves closer to the centre of enterprise infrastructure
Cloudflare said organisations increasingly need to secure AI across the full lifecycle, from employees using public AI tools to developers building applications and autonomous agents acting on behalf of companies.
Ercan Aydin, area vice president for the Middle East, Turkey and Africa at Cloudflare, said the move from AI pilots to production changes the infrastructure requirements facing organisations.
He highlighted the need for stronger visibility into AI traffic, governance controls around models and costs, and identity frameworks for AI agents.
The company is positioning its connectivity cloud as the underlying layer that brings together networking, security, developer services and AI infrastructure.
Cloudflare One and Workers form core of AI stack
Among the products being highlighted at LEAP is Cloudflare One, the company’s Zero Trust security platform for protecting users, applications and networks.
Cloudflare will also demonstrate Workers, its edge computing environment for deploying code, applications and AI workloads with low latency.
The company is positioning Workers as part of a broader model in which AI applications can be built and operated closer to users while avoiding some of the overhead associated with traditional infrastructure environments.
Firewall for AI targets prompt injection and data leakage
Cloudflare will also showcase Firewall for AI, which is designed to protect large language model applications against threats including prompt injection and data leakage.
As organisations deploy generative AI into customer-facing and internal applications, security controls around prompts, model behaviour and sensitive information are becoming increasingly important.
The company is also highlighting AI Gateway, which provides visibility and control over AI traffic, model usage and costs.
Cloudflare said a lack of visibility remains one of the biggest gaps for enterprises, particularly where employees use external AI services without central oversight.
Publishers gain more control over AI crawlers
The company will also promote AI Crawl Control and Pay Per Crawl.
These tools are intended to give publishers and enterprises greater control over how AI companies access and use their content.
The products reflect a wider debate around how generative AI systems collect, index and monetise online information, and whether content owners should have stronger mechanisms for restricting or charging for automated access.
Agentic AI brings new identity requirements
A major focus of Cloudflare’s LEAP presence will be infrastructure for autonomous AI agents.
The company will highlight isolated Sandboxes, Cloudflare Mesh, persistent Agent Memory, AI Search and native OAuth governance as part of its agentic AI portfolio.
Aydin said enterprises increasingly need to treat AI agents as users because they can log in, act and make decisions within corporate environments.
That creates new requirements around authentication, permissions and auditing.
An autonomous agent with excessive privileges and limited oversight could create a significant security risk, particularly when operating at machine speed.
Cloudflare links AI adoption with cyber threat acceleration
Cloudflare also expects AI to increase the speed of both innovation and cyberattacks.
The company referenced findings from its H1 2026 DDoS Threat Report, which reported a rise in attacks reaching terabit-per-second scale.
Cloudflare argues that AI will become increasingly important on the defensive side as security teams attempt to respond to threats that emerge and evolve faster than manual processes can handle.
The company sees the likely model as one where AI handles scale and speed while human security teams focus on judgement, strategy and exception handling.
Saudi Arabia becomes a key regional AI market
Cloudflare’s participation at LEAP reflects its continued focus on Saudi Arabia and the wider Middle East, Turkey and Africa region.
The company sees the Kingdom as an important market because of the pace of cloud adoption, AI investment and digital infrastructure development under Vision 2030.
At LEAP, Cloudflare plans to engage with government bodies, enterprises, technology partners and developers on how organisations can move AI workloads into production without losing control over security, identity and governance.
Why this matters
Cloudflare’s LEAP strategy reflects a broader shift in enterprise AI from model experimentation toward infrastructure management.
As AI becomes embedded in applications, workflows and autonomous agents, organisations need to manage not only the models themselves but also identities, permissions, data flows, costs and security controls.
The emergence of agentic AI makes this especially important because software agents are increasingly capable of taking actions inside business systems rather than simply generating responses.
Editor’s note
The most important theme is the convergence of AI infrastructure and identity.
Traditional security models were built around human users, applications and devices. Autonomous agents introduce a new category of actor that can authenticate, access systems and make decisions without continuous human intervention.
That means enterprises will need to extend Zero Trust principles beyond employees and workloads to machine identities. As Saudi organisations move AI into production, the ability to govern those non-human actors may become just as important as model performance itself.
