"60% of [NVIDIA's] revenue is still coming from the hyperscalers" (Ed Ludlow). "We have a full allocation to the tech stocks... the big hyperscalers" (Tony Roth). Public software companies are "presumed to be higher quality" and less susceptible to disruption than private ones (Laurie Goodman). The AI infrastructure buildout demand is still skyrocketing and constrained by compute supply. The primary customers and beneficiaries are the large, public cloud hyperscalers (Microsoft Azure, Google Cloud, Meta's AI, Amazon AWS). They have the capital, in-house talent, and scale to both drive and withstand AI disruption, unlike smaller, over-leveraged private software companies. They are the conduit for AI adoption. LONG on the dominant public hyperscalers. They are the capital goods providers for the AI cycle, have locked-in demand from NVIDIA's guidance, and are insulated from the private credit markdown crisis affecting their smaller competitors. Geopolitical disruption to tech supply chains (e.g., helium, sulfuric acid from the Middle East); eventual competition from in-house AI chips reducing reliance on NVIDIA; regulatory scrutiny.