Optimizing GPU utilization helps applications fully leverage their advanced computational power, which then leads to faster responses and smoother experiences. At the same time, AI is set to operate more independently, While both these scenarios can boost productivity, IT systems are at risk from increasingly sophisticated cyberattacks. Each of these areas is important for businesses to understanding and engage with.
To understand these developments further, Digital Journal heard from Richard Copeland, CEO of Leaseweb USA. This is in the form of three technology based predictions.
Prediction 1: Trusted Execution Environment Technology Will Reshape Distributed Compute and Multi-Cloud Architecture
An important area of strengthened cybersecurity is a Trusted Execution Environment. This is a segregated area of memory and CPU that is protected from the rest of the CPU by using encryption.
On this topic, Copeland foresees growth: “In 2026, Trusted Execution Environment (TEE) technologies will finally move from ‘interesting concept’ to real-world game changer. We’re going to see organizations secure memory and hardware in a way that simply wasn’t practical before, which opens the door for decentralized compute in a very big way. Companies will be able to safely split compute across multiple clouds, regional providers, and even on-prem environments, instead of keeping all their workloads under one hyperscaler’s roof. This will bring a level of flexibility and resilience that hasn’t been possible until now.”
As to the dynamics, Copeland assesses: “What is interesting to note here is that the shift isn’t driven by budgets or hype, but by behaviour. When you can secure workloads at the hardware level, you’re suddenly free to architect systems around business needs instead of who owns the data centre. It unlocks more creative architectures for blockchain, AI, and high-performance computing, and gives organizations confidence that they can spread their risk without compromising security.”
Prediction 2: AI Becomes Truly Agentic – Replacing Tasks, Not People – and Drives a New Phase of Cloud Repatriation
Agentic AI is an artificial intelligence system that can accomplish a specific goal with limited supervision. This is set to be an important businesses development area.
AI is shifting up a gear, observes Copeland: “AI is no longer just a tool for optimization. In 2026, agentic AI starts replacing full workflows, and that shift will separate companies that understand how to use AI from those that fight it. The real impact isn’t that AI replaces jobs, but that it replaces the tasks people shouldn’t be doing in the first place – the repetitive, time-sucking operations that drain teams. Organizations that lean into agentic AI will run faster, make decisions earlier, and redirect people into work that actually moves the business.”
There will be other developments: “As AI becomes more embedded in day-to-day operations, more companies will realize that complexity and cost are pushing them away from the hyperscalers. They’re seeing outages, noisy-neighbor issues, unpredictable billing, and environments so complex that one failure cascades through the whole stack.
AI workloads, especially GPU-heavy ones, run better, and more cost-effectively, when the infrastructure is simpler, more transparent, and built for their exact workloads. That’s why 2026 will be a major year for cloud repatriation back to regional providers and bare-metal platforms built for performance.”
Prediction 3: GPU Optimization and AI-Driven Attacks Will Push Companies Toward Regional Cloud Providers for Security and Stability
Getting more out of GPU will also be key, explains Copeland: “GPU optimization becomes a headline topic in 2026. Today, most companies only use about 60 percent of the GPU power for which they are paying. Next-gen optimization software is going to flip that on its head, giving organizations the ability to squeeze full value out of their infrastructure.
That matters not just for cost control, but for AI reliability. When your model performance becomes a competitive advantage, you can’t afford wasted compute, unpredictable throttling, or hardware carved into fractional units you can’t see. This is where optimized IaaS and regional GPU clouds start to shine.”
Copeland does warn, however, that cyberattacks are becoming increasingly complex and this needs to be taken into account: “At the same time, attackers are getting smarter, and they’re starting to use AI too. The largest, most complex cloud environments become the biggest targets – when bad actors can spin up their own LLMs.
Hyperscalers have hundreds of thousands of tenants, which means hundreds of thousands of potential attack surfaces (and pockets to pick). Regional providers have tighter vetting, cleaner environments, and fewer noisy neighbors. In 2026, security-conscious organizations will realize that the safest place to run AI and high-value workloads often isn’t the biggest cloud, it’s the one that actually keeps out the wrong people.”
