Commodity Watch: Stargate. Hyperion. Prometheus.

<div class="grid grid--33-66-col"><div class="col"><img loading="lazy" data-fr-image-pasted="true" src="/getContentAsset/e4db1c4c-2687-44cd-adbd-db1eb849e5d2/cb87803a-320c-480f-ab75-7b9029eaaf79/Ahmad-Al-Sati-new.jpg?language=en" alt="Ahmad Al Sati" title="Ahmad Al Sati" class="fr-fic fr-dii" style="width: 180px"></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>Stargate. Hyperion. Prometheus. These are not the names of a sci-fi trilogy but the names for some of the largest AI infrastructure projects currently being built for tens of billions of dollars. And this might only be the beginning. By some estimates AI infrastructure spend could exceed $5 trillion by 2030. This year alone, AI infrastructure spending by the largest players is projected at $320 billion. At ~1% of GDP, it exceeds the telecommunications infrastructure spend in the ‘90s, though not the 6% of GDP that railroads spent in the 1800s.</p><p>The capex is immense and most of these companies are betting that the ultimate returns from AI will be worth it. We will see the results at some point or not. But what is evident is that some of the largest tech companies are pivoting away from asset-light and low-capex business models where they designed the phone, but didn’t own the factory, or scaled software efficiently at low marginal cost (Instagram). Instead, AI infrastructure is causing some of the largest mega caps to move towards high-capex, asset-intense business models in the physical world (a resource intensive world is good for business- I am not complaining). As such, capex as a percentage of sales has increased from less than 15% to more than 25% over the last decade. At those rates, these companies resemble chip manufacturers more than software companies. That will impact margins and profitability, especially as depreciation begins to flow through the income statement.</p><p>The rules here are different as the benefits of scalability at lower marginal costs dissipate. The current spending on AI infrastructure is illustrative. To develop AI, companies are spending on three primary components: data centers (buildings), chips and power. All three require securing products from third parties (copper, steel, cement), all three are subject to scarcity constraints (scarcity of land, chips and energy), all three inputs require ongoing expenditures (electricity bills, upgrading hardware and maintaining a building for optimal PUE). Operational costs in the real world are also higher. McKinsey estimates that the labor costs for AI infrastructure will be $500 billion.</p><p>The benefits from AI may ultimately outweigh the costs. But buildouts are not seamless. Risks and chokepoints remain. As everyone who has supervised a construction project or remodeled a kitchen knows, stuff happens. Where will the materials necessary to build tens of millions of square feet come from? How will we secure the requisite energy equivalent to powering 150 million homes? How will companies accustomed to dealing with bits and bytes deal with exogenous shocks? Meta’s Hyperion will be built in Louisiana (Hurricanes anybody?). Rather than digitizing the world further, AI seems (at least in this stage) to be moving us into a world reliant on scarce resources and global supply chains (chips). To say it one more time, supply chains matter.</p><p>PS: Not AI content. Not investment advice.</p><p><img loading="lazy" src="/getContentAsset/65570ba9-1403-4025-8d7c-0bd3c4a08d30/cb87803a-320c-480f-ab75-7b9029eaaf79/graph.jpg?language=en" alt="Global data center capital driven by AI graph" title="Global data center capital driven by AI graph" style="width: 100%" class="fr-fic fr-dib"></p><p><br></p>