UBS maintains $1.2 trillion AI capital expenditure forecast despite safety calls

The decoupling of infrastructure spending from frontier model breakthroughs suggests the AI investment cycle has moved into a more mature phase. For streaming and media companies, this indicates that the cost of running AI-driven recommendation engines and generative tools will remain high, even if the pace of foundational model releases slows. The shift toward inference-heavy demand means hardware providers like memory and networking firms will see sustained revenue regardless of regulatory hurdles for new models. As safety standards become formalized, larger incumbents may gain a competitive advantage by absorbing the costs of independent audits. Watch for the impact of US midterm election rhetoric on data center energy regulations to see if physical constraints begin to limit this projected spending.