Investors backing ‘wrong future’ in AI, strategist says — new winners emerge (www.cnbc.com)

Investors are currently pouring resources into large AI data centers, but Joachim Klement, managing director at Panmure Liberum, warns they may be backing the wrong technological future. He predicts a shift towards small language models capable of running on consumer devices, such as desktop computers and mobile phones, rather than relying solely on extensive data center infrastructures for AI processing. Citing Stanford research, Klement notes that these smaller models can effectively handle over 80% of typical AI tasks, suggesting that major beneficiaries of this trend will be device manufacturers like Apple and Dell, rather than the hyperscalers currently dominating the AI landscape. This potential pivot in the AI landscape could dramatically alter investment strategies and operational approaches for businesses. Companies are already becoming judicious about how they deploy AI, matching tasks to models based on cost and performance. Klement highlights that as businesses adopt local processing, the economics become compelling—initial capital expenses and electricity costs for local models are significantly lower than those associated with data centers. As this landscape evolves, companies in edge computing, such as Arm and Qualcomm, are poised to gain from the shift away from hyperscaler dependence, indicating a broader reconfiguration of the AI ecosystem in favor of more decentralized computing solutions.