Most GenAI models are developed and implemented in substantial datacenter clusters, but creating, testing, and prototyping locally remains crucial today. Traditionally, this required high-end, costly workstations. Nvidia’s DGX Spark, despite its relatively modest power, offers a 128GB AI lab in a compact form, enabling versatile AI workloads. However, it’s not the only choice. AMD’s ‘Strix Halo’ stands out, offering similar functionality at a lower cost, with software stacking akin to their datacenter offerings, easing migration to larger setups. Testing with HP’s Z2 Mini G1a, both systems proved capable across AI tasks, from single-user processing to fine-tuning and image creation. AMD’s architecture supports essential AI development while Nvidia focuses on network capabilities and high-speed operations ideal for AI environments, determining consumer preference based on priority, be it GPU power or comprehensive processing capability. Performance metrics highlight the strengths and compromises of each system, ensuring IT professionals identify the most suitable option for specific AI demands.