AI AND DATA CENTRES
Keeping up with demand AI is rapidly changing how businesses work, but it is pushing IT infrastructure beyond its traditional limits, presenting new challenges for the businesses that rely on data centres. For channel businesses, this creates opportunities to offer more sophisticated solutions and become strategic advisors in AI adoption.
While AI adoption continues to grow at pace with owners looking to harness its potential to add value to the business, the compute power it requires is pushing most existing data centre infrastructure beyond the workloads it was designed for. “AI demand is now moving faster than infrastructure can keep up, creating bottlenecks around power, cooling, compute capacity and networking that can ultimately limit how quickly organisations can put AI into production,” says Steve Jackson, field CTO at SHI. “That can then threaten the viability of AI projects particularly given how quickly the sector is moving, with the technology making giant leaps in terms of its capabilities every few months. “For a business looking to roll out or scale its AI capabilities, it's the data centre that is increasingly thwarting those ambitions, so organisations need to understand where genuine constraints exist and work backwards from the outcomes and workloads they need to support, assessing their AI readiness and developing an infrastructure strategy that is achievable, sustainable and focused on genuine business value.” Ian Stretton, presales consultant at Infinity Group, notes that the danger for businesses is investing heavily in AI without looking at the wider architecture. “A powerful GPU estate achieves little if data cannot move quickly enough between storage, memory and processors,” he explains. “Bottlenecks can leave expensive compute sitting idle, increase processing times and ultimately push up the cost of AI. AI performance depends on the balance between compute and data transfer, rather than compute alone.”
42
Powered by FlippingBook