Speaking from Davos, Nvidia’s CEO described the ongoing expansion of AI computing as “the largest infrastructure buildout in human history,” with global spending on data centers, chip fabs and so-called “AI factories” forecast to reach $85 trillion over the next 15 years. That projection encompasses land acquisition, power distribution and cloud-scale server deployments. For investors, this implies sustained multi-year growth in capital expenditures across hyperscalers, colocation providers and semiconductor equipment suppliers serving Nvidia’s accelerated GPUs.

Nvidia reports that spot rental rates for its GPUs are climbing steadily as enterprise and startup R&D budgets shift heavily toward AI workloads. To address cost pressures, the company’s latest H200 series delivers double-digit improvements in energy efficiency and up to 30% lower token-processing costs compared with prior generations. Those efficiency gains should broaden end-market adoption, extend total addressable market and support higher overall revenue without sacrificing gross margins.

Following formal approval by U.S. regulators, Nvidia resumed shipments of high-end AI accelerators to China and other countries. Management has downplayed military dependency risks, noting that local foundries and domestic chip designers supply their own defense customers. The resumption of exports reopens a critical revenue stream projected to contribute a mid-single-digit percentage of total sales in the coming quarters, supporting analysts’ consensus for high-teens year-over-year revenue growth.

Beyond organic GPU sales, Nvidia has deployed $150 million in a $300 million funding round for AI inference startup Baseten, while committing up to $100 billion in compute credits to partnership with leading generative-AI developers. Dozens of minority investments in firms building application frameworks and physical-AI robotics platforms further entrench Nvidia’s technology as the foundational layer of the next industrial revolution. For investors, these stakes offer optionality on software-as-a-service margins and potential future M&A upside.