Africa is scaling AI, not experimenting, says Dell exec
Africa and its CEEMETA peers are moving from pilots to production at remarkable speed, according to Mohammed Amin, regional executive at Dell Technologies. He argues that the companies building and deploying AI fastest—on top of intelligent, fit-for-purpose infrastructure—will set the pace for the region’s digital economy.
Speed is the new competitive edge
Amin says AI has reset expectations for how quickly work gets done. Hyper-automation is replacing manual, multi-step workflows with processes that complete in seconds. Government departments in the Gulf are using AI to streamline citizen services; contact centers in Poland and South Africa are cutting resolution times with generative AI and real-time data; even complex functions such as financial modeling are now executed almost instantly. Countries including the UAE, Saudi Arabia, Poland and Turkey have accelerated national AI strategies over the last 18 months, and enterprise adoption is following suit.
Infrastructure under pressure—and the rise of hybrid
Much of today’s enterprise stack was not built for modern AI and cloud-native patterns. The surge in unstructured data—growing at roughly 55% annually—exposes that gap. Amin expects 2026 to be a decisive year for hybrid strategies that match workloads to the right execution environment.
His view: critical data and high-value AI workloads will increasingly stay on-premises to maximize security, control costs and meet data sovereignty requirements, while public cloud delivers elasticity for less sensitive tasks. AI PCs and edge systems will push inference closer to where data is created, reducing latency and keeping sensitive information local. The payoff is optionality—teams can choose infrastructure based on performance, cost and control, rather than a one-size-fits-all model.
Tokens are the new workload metric
Amin highlights a rapid surge in “token consumption”—the unit cost of every AI interaction—as a defining stress test for enterprise tech stacks. In financial hubs like the UAE, a single AI request can now trigger dozens of downstream actions touching APIs, identity and compliance checks, databases and more. Each step consumes tokens, and as volumes spike, they expose bottlenecks across storage, networking, compute, security and governance.
“Token growth is outpacing raw compute,” he notes, arguing that success demands more than advanced GPUs. High-bandwidth networking, robust storage tiers and seamless orchestration are now essential to unlock the full performance of agentic and generative AI at scale. Organisations that tune the entire stack—not just accelerators—will deliver more reliable, secure and scalable AI experiences.
From centralized AI to intelligence everywhere
The region is shifting from monolithic, centralized AI to distributed intelligence that blends large foundation models with micro LLMs. The smaller models consume far less power—a crucial advantage as energy costs rise—and can run at the edge where data is born. That opens new possibilities for sectors operating in remote or bandwidth-constrained environments, from African mining sites to Polish manufacturing lines. In this model, big models anchor strategy and complex reasoning, while micro models deliver fast, local decisions.
AI and cybersecurity converge
With CEEMETA among the world’s fastest-growing targets for cyberattacks—and threat actors now weaponizing generative AI—security is inseparable from AI deployment. Amin expects organisations to double down on zero-trust architectures, rigorous data governance and hybrid environments that keep sensitive AI workloads protected. AI PCs and edge devices will act as frontline defenders, detecting and mitigating threats locally while feeding insights back into central security operations.
Physical AI: from code to the real world
Amin sees the next wave arriving in robotics. Instead of hard-coding machines for narrow tasks, enterprises can now set goals and let AI-enabled robots learn through experience. This “physical AI” is moving beyond factory floors to tackle repetitive or dangerous work—from hazardous inspections to high-variability logistics. Companies that deploy purpose-built, AI-driven robotics will operate at a speed and scale that traditional automation can’t match.
The 2026 mandate
Across Central and Eastern Europe, the Middle East, Turkey and Africa, the race to industrialize AI is on. The winners, Amin says, will move fast but responsibly: embedding AI into core operations, optimising the full stack (compute, storage, networking, security and orchestration) and adopting hybrid architectures that balance control with scale. For Africa in particular, the message is clear—this is no longer experimentation. It’s an execution play, powered by intelligent infrastructure and measured by real outcomes in government services, customer experience and industry productivity.
“AI has become the new performance baseline,” Amin concludes. “Those who build quickly, secure intelligently and scale responsibly will define the region’s digital economy.”
