Commodity Watch: The New Power Struggle

<div class="grid grid--33-66-col"><div class="col"><img loading="lazy" src="/getContentAsset/061c994a-a452-418f-bfa0-f2cf3cf5c577/cb87803a-320c-480f-ab75-7b9029eaaf79/Ahmad-Al-Sati-new.png?language=en" alt="Ahmad Al Sati - insights" title="Ahmad Al Sati - insights" style="width: 180px" class="fr-fic fr-dii"></div><span style="font-size: 12px"><div class="col"><strong>AHMAD AL-SATI<br></strong>PORTFOLIO MANAGER<br><br><p>Ahmad is the President and Portfolio Manager for Gemcorp Capital Advisors LLC, based in New York.&nbsp;<br><br>Ahmad has spent most of his career in the global credit markets. Prior to Gemcorp, Ahmad was President of Pandion Mine Finance and RiverMet Resource Capital, LP - a fund focused on investing in precious metals, where he was responsible for managing the investments and the day-to-day operations of the registered investment adviser.&nbsp;</p></div></span></div><hr><p>The struggle for power is here. Not economic or political power. But electric power. Electricity is integral for both our modern lives and for the seemingly defining technology of this time, artificial intelligence (AI). Last week, Alphabet announced that it is acquiring Intersect Power for approximately $4.75 billion. Intersect develops large-scale power projects combining renewables, batteries and natural gas to create micro grids dedicated to data centers. The acquisition comes a year after Alphabet partnered with Intersect to develop around $20 billion of power generation projects for its data centers. Now that expertise and accompanying pipeline is going in-house.</p><p>Alphabet is not alone. Meta, Amazon and Microsoft are all pursuing long term power purchase agreements as they source electricity from hydrocarbons, renewables and nuclear because it is increasingly evident that a major bottleneck to AI growth is access to reliable, scalable and continuous electricity.</p><p>AI data centers use a lot of electricity (new GPUs use 3x more watts than CPUs). In Ireland, data centers are now consuming 21% of national usage, forcing a moratorium on new data centers. In the US, data centers are using 4% of the US’s electric usage. More importantly, data centers are increasingly competing with consumers for electricity resulting in Not In My Back Yard (NIMBY) resistance.</p><p>A solution: bring your own power using all available sources of electricity. Dedicated micro grids developed concurrently with data centers allows the hyperscalers to reliably project power availability. Significant demand also means that data centers cannot be choosy. They are and must utilise all potential sources of electricity from solar to wind to nuclear.</p><p>The resurgence of physical materials thus continues. Power development and production require actual materials, and it obeys different scalability laws compared to software. Increased reliance on physical assets requires different approaches, skills and capital (an economic paradigm shift in the works perhaps). For physical assets constant investments are required for constant growth. More capex is tied to more power generation in a way that it is not for software as a service (SaaS) companies. You can’t just build one power plant and scale it ad infinitum. In turn, that means increased demand for the materials necessary to produce electricity such as minerals and metals in batteries, solar panels and turbines as well as feedstock such as natural gas.</p><p>Securing these building blocks remains subject to complex supply chains. And even as these supply chains matter more, they are increasingly fragile because of geopolitics, processing concentration, and under investment. Chips were the first bottleneck for AI growth. Now, power is the second bottleneck. Yet, the power chokepoint ultimately requires more investment in the commodities necessary for electricity. Producing, processing, trading and transporting these increasingly essential materials may therefore come to the fore during an AI age in a way they didn’t when software ate the world.</p>