While Oracle Will Rake in Big Bucks on AI, Profits Are Hard to Predict (www.nextplatform.com)

Oracle is undergoing a monumental transformation as it pivots to become the fifth major U.S. hyperscaler, investing heavily in infrastructure with plans for five new datacenters to bolster its AI processing capabilities. The company is ramping up its datacenter capacity, with 1.2 gigawatts launched in fiscal 2026 and another gigawatt set to come online shortly. This infrastructure expansion is part of Oracle’s strategy to tap into an anticipated revenue backlog of $638 billion, which includes significant contributions from key customer OpenAI. CFO Hilary Maxson indicated that Oracle expects to recognize around $76.6 billion in revenue in fiscal 2027 from this backlog, suggesting robust growth opportunities. Despite the substantial capital investments—$291.6 billion over three years—Oracle's management believes that selling excess GPU capacity will help mitigate risks associated with reliance on OpenAI, given the surging demand for GPU resources in AI applications. With GPU utilization reportedly at 97.5%, Oracle is strategically positioned to capitalize on high demand, even if some customers churn. The company’s recent financial results reflected strong growth, particularly in its Cloud Infrastructure division, which saw a remarkable 93.2% increase, fueled by the booming market for AI and cloud services. This strategic shift heralds significant implications for the AI/ML community, as Oracle aims to cement its role as a leading provider of AI infrastructure.