From Gigawatts to Tools: Sizing the AI Build-Out's Demand for Wafer Fab Equipment

11:10 am - 11:35 am

The AI infrastructure build-out is typically measured in gigawatts of compute that are set to be deployed over the coming decade. For the equipment, materials and manufacturing industry, the task is translating those headline gigawatt figures into a realistic plan for the wafer capacity, equipment and tooling the build-out requires.

 

This session presents a bottom-up framework that traces AI demand from gigawatts of deployed compute, through accelerators, their memory and logic content, and the wafer capacity they require, down to wafer fab equipment spending by segment. 

 

It starts by separating announced gigawatts from the capacity that can realistically be energized and brought online, using a bottom-up data-center model, then sizing the bottlenecks that determine how quickly that demand can be met. The aim is to give capacity decision-makers a demand-anchored way to stress-test their own assumptions: where AI demand converts into durable tool demand, and where the real bottlenecks lie, from advanced packaging and DRAM capacity to EUV.

 

Key Technologies Covered

  • GWs to wafer fab equipment (WFE)
  • AI accelerator and custom-silicon (GPU/ASIC)
  • HBM and DRAM
  • Advanced packaging (CoWoS / 2.5D-3D)
  • EUV lithography
  • Leading-edge logic

Featured Speakers

Mr. Rolf Bulk

Mr. Rolf Bulk

Head of Semiconductors & Infrastructure, Futurum Group

Rolf Bulk is Head of Semiconductors and Infrastructure at Futurum Group, where he leads the team covering the global semiconductor and AI-infrastructure value chain. His research focuses on the AI compute stack, spanning data-center power demand, accelerators and custom silicon, logic and memory semiconductors, foundry, advanced packaging, and semiconductor capital equipment.

His work is centered on Futurum’s proprietary framework, which connects data-center power demand to accelerator deployments, memory and logic content, wafer capacity, equipment spending, and company-level revenue and earnings implications. The framework helps institutional investors assess where AI infrastructure demand is translating into durable semiconductor growth, where supply-chain bottlenecks are emerging, and where market expectations may prove too conservative or too aggressive.

Before joining Futurum, Rolf was a Senior Analyst and co-founder of the Technology vertical at New Street Research, where he covered semiconductor and infrastructure stocks across Europe, the US, and Asia. Earlier, he was part of Bernstein’s European semiconductor team, which held an Institutional Investor #1 ranking. He holds an MSc in Finance from the Stockholm School of Economics and is based in Singapore.