AI is often discussed in terms of the large computing environments needed to train and run advanced models. But its business value also depends on what happens closer to employees: whether they can use AI tools to find information, analyse data, develop content and carry out their work effectively.
That means looking beyond the data centre. AI workloads can run across large clusters, cloud and neocloud environments, data platforms, edge locations and employee devices. Each has different requirements, and organisations need to plan how these elements support the use cases they want to deliver.
“The AI infrastructure conversation does not stop at the data centre. It stretches from the largest AI clusters to the device on every desk,” says Marc Layne, Sales Director, Dell Technologies South Africa.
“Training, inference and everyday employee productivity place different demands on infrastructure. Organisations need to think across that continuum so that the capabilities being built centrally can translate into useful, secure experiences for people at work.”
Connecting infrastructure to business use
When AI infrastructure, data strategy and workplace modernisation are planned separately, gaps can slow adoption. A business might invest in computing capacity without preparing its data for AI, or introduce applications while employees lack suitable devices, secure access or the skills to use them.
A connected approach starts with the business problem. Teams can then assess the computing, data, security and user experience needed to address it. Some workloads require high-performance computing and storage; others may benefit from running closer to where data is generated or used. Employees also need appropriate devices and guidance on using AI responsibly.
“The objective is to align infrastructure choices with the outcomes an organisation wants to achieve,” says Layne. “That means considering where workloads should run, how data will be protected and governed, and what employees need to use AI effectively in their day-to-day roles.”
The role of neoclouds and AI at scale
The growth of AI-focused infrastructure providers, including neoclouds, highlights the specialised requirements of AI at scale. Training and inference can place significant demands on computing, storage, networking, cooling and operational services. Adding capacity therefore requires more than servers: the components need to work together and be managed as demand changes.
For organisations building or using these environments, performance needs to be weighed alongside reliability, security, cost and operational control. Infrastructure must also connect effectively to the data and applications that turn model capabilities into useful services.
Modern devices bring AI closer to employees
At the other end of the continuum, AI PCs and modern endpoints are becoming part of workplace planning. They can support AI-enabled applications and, where suitable software and hardware are available, run certain AI tasks locally. The right approach depends on the task, the sensitivity of the data, the applications available and the organisation’s security requirements.
Dell’s EMEA research indicates that 86% of organisations see AI PCs and modern endpoints as important to their future workplace strategy. In the same regional findings, 79% of leaders say outdated devices limit adoption of new technologies, including AI-powered applications, while 92% say modern devices improve collaboration and workforce productivity. These findings cover EMEA and are not specific to South Africa.
For South African businesses, the practical question is whether employees have the tools and support to use AI safely and productively. Devices are one part of the answer; organisations also need policies, training and workflows that make the technology relevant to people’s roles.
“AI value becomes real when people can apply it to everyday work, whether that means finding information more quickly, collaborating more effectively or using role-specific assistants,” Layne says. “The device layer matters, but it has to be part of a wider plan that includes security, data governance and skills.”
Planning across the full continuum
A connected AI strategy considers large clusters, cloud and neocloud environments, enterprise data platforms, edge systems and employee devices together. This helps organisations match each workload to an appropriate environment and understand how the parts will support one another.
A practical starting point is to identify measurable business needs, assess infrastructure and data readiness, address the most significant gaps and help employees develop the skills to use AI responsibly. Dell’s portfolio spans AI infrastructure, storage, services and AI PCs, helping customers plan across these environments.
As AI becomes more distributed, organisations need a strategy that connects investment in infrastructure to useful outcomes at the point of work.
©Higher Education Media Services.