Energy Transition

Hyperscaler Expansion | From Growth to Grind

byAdam Robinson, Enverus Intelligence® | Research (EIR) Contributor

Hyperscalers’ capital expenditure forecasts show a split between expanding and maintaining data center (DC) capacity, with upkeep costs quickly outpacing growth spending. Computer chips on a five-year replacement schedule cost about $27 billion/GW of IT at the leading edge. Enverus Intelligence® Research’s new capacity expansion report analyzes trends in DC development in the L48, projecting capacity buildout over the next decade.

We predict the increase in power capacity, coupled with chip performance improvements will lead to a 33x surge in total compute by 2035. The results of this rise can be seen through efficiency and performance gains in various large language models. Meta’s Llama 3.1 seventy billion (70B) achieved 83% on the massive multitask language understanding (MMLU) benchmark, at a 99% discount compared to OpenAI’s GPT-4. Soon after, DeepSeek’s R1 model cut costs by 97% within only two months of GPT-o1’s release, achieving higher than 90% on the MMLU benchmark.

Research Highlights

  • Long-Term Load Forecast – The Data Center Decade : This report encompasses Enverus Intelligence® Research’s view on how exponential load drivers will impact our power demand forecasts from 2025-2050.
  • Hyperscaler Expansion – Stepping Onto the Capex Treadmill : This report provides a refreshed L48-level view on expected growth in data center capacity and associated power demand with a focus on understanding hyperscaler capital programs.
  • Reshoring Revival – Powering U.S. Industry : This report examines the impact of manufacturing load resulting from increased domestic manufacturing driven by demand, tariffs and production economics. We also explore regional load growth variations influenced by infrastructure and adoption trends, the impact of recent legislation and tax credits on manufacturing, and the comparative economics of imports versus domestic production.

Enverus Intelligence® | Research, Inc. (EIR) is a subsidiary of Enverus that publishes energy-sector research focused on the oil, natural gas, power and renewable industries. EIR publishes reports including asset and company valuations, resource assessments, technical evaluations and macro-economic forecasts and helps make intelligent connections for energy industry participants, service companies and capital providers worldwide. See additional disclosures here.

Picture of Adam Robinson, Enverus Intelligence® | Research (EIR) Contributor

Adam Robinson, Enverus Intelligence® | Research (EIR) Contributor

Adam Robinson joined the Enverus Intelligence® Research team as an Energy Transition Analyst in October 2024, focusing on supporting sustainable energy solutions. He holds a degree in Chemical Engineering from the University of Calgary and brings hands-on experience in pipeline integrity engineering. With a strong foundation in pipeline systems and integrity management, Adam is passionate about leveraging data-driven insights to optimize energy infrastructure and enhance system reliability.

Subscribe to the Enverus Blog

A weekly update on the latest “no-fluff” insight and analysis of the energy industry.

Related Content

Blackstone announces agreement to acquire Enverus
Energy Analytics
ByAmyra Mardhani, Enverus Intelligence® | Research (EIR) Contributor

Gevo recently announced the sale of carbon dioxide removal (CDR) credits from its North Dakota facility for the first time, following the acquisition of an ethanol plant with carbon capture capabilities nearly a year ago. This is one of several...

Enverus Intelligence® Research Press Release - Waha prices expected to go negative (again)
Midstream
ByEnverus

Midstream is caught between a rock and a hard place. On one hand, gathering and pipeline operators, processors, storage and energy marketers face abundant opportunities and challenges when meeting customer demand.

Enverus Press Release - Speed through records with Enverus Instant Analyst™ - Courthouse
Midstream
ByDaniel Steffy

Let’s be honest:How many spreadsheets are you using to forecast production? If the answer is more than one, you’re not alone — and you might be stuck in what many teams experience as spreadsheet chaos. Every team has their own...

  • 2
  • 3