Why the search for “peak AI” spending is misguided VIEW IN BROWSER Imagine the scene. A major CEO calls a surprise press conference. Cameras flash. Reporters crowd forward. The room goes quiet as he steps to the podium. “We’re done investing in electricity,” he announces. A stunned, confused silence. “We’ve reached peak power usage. From here on out, we won’t be spending another dollar on it.” The statement is so absurd that it barely deserves a follow-up question. No serious business leader would ever say such a thing. Electricity isn’t a one-time project. It’s a utility. The more you produce, the more you grow, the more customers you serve — the more power you consume. Usage scales with growth. And yet, many investors think AI is a one-time spend. By the numbers, global AI spending is projected to exceed $2.5 trillion in 2026 alone. Wall Street keeps looking at this and asking: Have we reached peak AI spending? That question only makes sense if AI is a one-time project – something companies build, finish, and then move on from. Yes, infrastructure such as data centers will be built and finished. But AI is not just data centers. It’s an entire ecosystem of infrastructure – chips, networking, cooling, power, storage, software, and security – that demands ongoing investment just to maintain performance, and even more spending to keep up with the next generation of models. It’s becoming a utility – an always-on, metered layer of digital infrastructure that businesses will draw from every minute of every day. This is a structural shift, not a cyclical spike. Utilities don’t have “peak build” moments. They expand with demand. And if AI is a utility, then the idea of “peak AI” spending is premature. So, if AI were a one-and-done project, it might make sense to ask whether spending is peaking. But it’s not, so the question itself is fundamentally flawed. AI spending isn’t ending. It’s evolving. And investors who mistake evolution for exhaustion are putting themselves on the wrong side of history’s most significant infrastructure buildout. How You Build a Utility The idea of, and fretting over, “peak AI” spending is everywhere in the mainstream media. From MarketWatch:
From CNBC:
The numbers are staggering. Just five companies – Amazon, Microsoft, Google, Meta, and Oracle — are expected to spend over $600 billion on AI infrastructure in 2026. Huge spending numbers like that should make everyone take notice. But that’s why it’s so important to change your thinking. Utilities don’t just require software. They require infrastructure. Think for a moment about everything needed for massive, durable, national-scale infrastructure. - Electric grids
- Power plants
- Semiconductor fabrication
- Rare earth processing facilities
- Data centers
- Energy pipelines
Utilities aren’t software upgrades. They are physical ecosystems. If AI is going to fulfill its potential and change everything about the way we live and work, then someone must secure the entire supply chain behind it. That’s where President Donald Trump and the U.S. government come in. Too many investors are missing this part of the story. While Wall Street debates whether AI spending has peaked… Washington is preparing for something much bigger. We’re not just talking about tax breaks or breaking down regulatory hurdles. It’s not just funding research. It’s taking stakes in companies that will provide the buildout needed for this new utility feature of American Life. Because if AI is a utility, then controlling the supply chain behind it becomes a matter of national security. That means directly backing, with cash, the companies that can do the building, supply the raw materials, bring the semiconductors to market, and guarantee energy to make it all work. One of the few analysts who has been connecting these dots is Luke Lango. In his new research for subscribers, he lays out what he believes is the next phase of this story – a second wave of government-backed investments that go far beyond pouring dollars into a few rare earth miners. According to Luke, this isn’t just about stockpiling minerals; it's about rebuilding the entire “mines-to-magnets” supply chain. That includes everything from raw materials… to processing… to semiconductor fabrication… to AI data center hardware… to the energy grid that powers it all. In other words, the full stack behind the AI utility. Here is how Luke wrote about it to his Innovation Investor subscribers: Last year, the government funneled more than $11.3 trillion … between direct spending, subsidies, and industrial incentives … into rebuilding America’s strategic backbone. And according to senior officials, that number is set to grow as Washington quietly prepares for what insiders describe as a long, grinding technological arms race. Here’s the part almost no one understands… The biggest winners won’t be the obvious ones. Not Big Tech. Not the trillion-dollar giants in your index fund. But the small, overlooked companies being pulled—sometimes shoved—into Washington’s inner circle. The ones that make the chips behind the chips… the power behind the power… the materials behind the machines. If history teaches anything, it’s that when Washington rewires the economy, fortunes shift behind the scenes. It happens quietly at first, then violently all at once. For investors, the opportunity isn’t in the companies the government has already backed, but in the companies it may need to back next. That’s the thesis behind his new report, The President’s Market — where he identifies five specific stocks he believes sit at the chokepoints of this supply chain shift. AI isn’t a one-time buildout. It’s becoming as foundational as electricity. And when a technology reaches that level, spending doesn’t peak… it compounds. Forecasts suggest AI infrastructure investment could grow toward $758 billion by 2029, with total global data-center related spending approaching $7 trillion by 2030. The question isn’t whether AI investment is slowing. It’s whether you’re positioned for the next phase of the buildout. Enjoy your weekend, Luis Hernandez
Editor in Chief, InvestorPlace | 
No comments:
Post a Comment