Google Deepmind CEO says the memory shortage is creating an AI 'choke point' (www.businessinsider.com)

Google DeepMind CEO Demis Hassabis has highlighted a significant "choke point" in the AI industry due to a severe shortage of memory chips, stating that the entire supply chain for such components is heavily constrained. Despite Google manufacturing its own Tensor Processing Units (TPUs), Hassabis mentioned that critical components remain in short supply, impacting not only their production capabilities but also the broader AI research landscape. The high demand from AI companies, coupled with escalating costs and limited availability, has intensified competition for these essential chips. The implications of this memory shortage are substantial for the AI/ML community. It restricts the ability to experiment with new innovations at scale, which is crucial for advancing AI models like Google's Gemini. This challenge is exacerbated by the predominant reliance on just a few suppliers—Samsung, Micron, and SK Hynix—who are struggling to cater to AI hyperscalers while maintaining service levels for traditional electronics customers. As AI firms like Google forecast significant capital expenditures in the coming years, the memory chip crisis poses an ongoing challenge that may slow the pace of AI development and deployment.