Businesses are struggling to achieve AI success - this report reveals why (www.techradar.com)

A recent report by DDN highlights that many businesses are facing significant challenges in their AI initiatives, with over half canceling projects due to inadequate infrastructure. The survey reveals that 65% of enterprises find their AI environments overly complex to manage, and 98% acknowledge skills gaps within their teams. The report emphasizes that most failures are rooted in siloed systems in storage, compute, and data pipelines, indicating that the issue is less about computing power and more about effective integration. Notably, 97% of respondents believe that adopting cloud solutions is crucial to improving scalability, with expectations for hybrid AI workloads to surge by 162% in the coming year. Moreover, the growing scrutiny of AI's energy efficiency reveals another layer of complexity. A significant 93% of companies are under pressure to minimize AI's energy impact, with many pointing to power and cooling inefficiencies as major concerns. This shift has brought forth the concept of "tokens per watt" as a new metric for evaluating AI performance, underscoring the urgent need for unified, modern infrastructure to support AI aspirations effectively. Companies are being urged to prioritize data management over simply acquiring advanced models and GPUs to truly harness the potential of AI technologies.