← Back to Library

Thank God for data centers

Packy McCormick challenges the prevailing narrative that AI data centers are merely wasteful energy sinks, arguing instead that they are the unlikely engines of American reindustrialization. In a piece that reframes a controversial infrastructure boom as a providential catalyst for hard tech, McCormick suggests that the insatiable demand of these facilities is doing what government procurement once did: funding technologies before they are ready, forcing them down the learning curve, and making them viable for the rest of the world.

The Meta-Alpha Product

The core of McCormick's argument rests on a historical pattern: new technologies often fail because they are too expensive to compete with incumbents until a specific, high-value customer forces them to scale. He identifies the data center not just as a consumer of chips, but as a unique economic actor. "Data Centers are increasingly serving as Buyers of Capabilities, acting as something between a government and a commercial buyer," McCormick writes. This distinction is crucial; unlike a typical commercial buyer who shops for the lowest price, these facilities need capability now, regardless of cost, to meet the demands of AI development.

Thank God for data centers

McCormick draws a parallel to the "Alpha Products" of the past—like the Sony Handycam for lithium-ion batteries or the calculator for microcontrollers—but notes that data centers operate on a vastly larger scale. "If you can sell them something they need, fast, they have an almost bottomless bid," he observes. This creates a rare environment where companies building advanced nuclear reactors, enhanced geothermal systems, or high-voltage direct current grids can secure revenue without waiting for the market to mature naturally. The author posits that this dynamic is "a commercial analog operating on DoD-style procurement logic but commercial timescales."

This framing is compelling because it shifts the focus from the controversial output (AI models) to the necessary inputs (physical infrastructure). It suggests that even if the AI bubble bursts, the physical advancements in energy and construction funded by this boom will remain. "Data Centers are funding the future where no one else will," McCormick asserts, a line that captures the urgency of the situation. However, critics might note that this optimism assumes the capital is being directed toward genuinely novel technologies rather than simply inflating the cost of existing, inefficient solutions. The risk is that the "bottomless bid" could sustain a bubble of overcapacity rather than true innovation.

Far from being the villains they are painted as, Data Centers may be the greatest accelerant of American Reindustrialization and a built-world future that benefits all people that we've ever seen.

The Apollo Parallel

To contextualize the public backlash against these massive energy consumers, McCormick reaches back to the Apollo program. He reminds readers that the space race was deeply unpopular at the time, with polls showing a majority of Americans opposed to the cost and prioritization of lunar missions while problems on Earth remained unsolved. "President John F. Kennedy gave his canonical 'We choose to go to the Moon' speech not because it was popular, but because it was unpopular and he needed to rally support," McCormick notes. The historical record shows that even a decade after the moon landing, public support for the program's costs was still divided.

The author uses this history to illustrate a recurring theme: the immediate costs of ambitious projects are visible and painful, while the long-term technological dividends are invisible until they arrive. Just as the Apollo program accelerated the development of integrated circuits, fireproof fabrics, and water filtration systems, the current data center boom is doing the heavy lifting for technologies like solid-state transformers and modular construction. McCormick points out that "the absurdity of the task's ambition coupled with the bottomlessness of its budget... were exactly the conditions needed to create terrestrially-useful innovations that would otherwise have taken much longer, or never been invented at all."

This historical lens adds significant weight to the argument, suggesting that the current friction is a necessary growing pain. The comparison to the "Mother of All Demos" and the funding of early AI labs by the Defense Advanced Research Projects Agency (DARPA) further cements the idea that high-stakes, high-cost procurement has always been the engine of general-purpose technologies. Yet, the analogy is not perfect; unlike the centralized, mission-driven nature of the Cold War space race, the current data center boom is driven by private competition and profit motives, which may lead to different outcomes in terms of equity and distribution.

The Learning Curve Advantage

Perhaps the most distinct contribution of the piece is the focus on the "learning curve." McCormick explains that many superior technologies, such as advanced nuclear reactors, are currently too expensive because they lack scale. "The challenge with these technologies, in normal times, is that there is little economic incentive for the buyers who would enable the scale to stick their necks out," he writes. Natural gas is cheap and abundant, creating a stalemate where the superior technology cannot break through.

Data centers break this stalemate by providing the initial demand needed to drive costs down. "They offer dilution-free capital (real revenue on a negative working capital cycle) to fund the big vision, and more importantly, the opportunity to get to scale and down the learning curve years earlier than would otherwise have been possible," McCormick argues. This transforms the data center from a passive consumer of electricity into an active investor in the industrial base. The author suggests that this mechanism "makes companies that might otherwise have died in the Valley of Death viable."

The argument holds up well against the backdrop of current supply chain constraints, where developers are indeed willing to pay premiums for speed and reliability. However, the piece glosses over the potential for resource misallocation. If the demand for data centers is temporary or if the technology stack shifts rapidly, the specialized infrastructure built for them could become stranded assets. The assumption that the "learning curve" will inevitably lead to cheaper, better technologies for the broader market is a strong bet, but it relies on the continuity of this specific, high-cost demand.

In five years, this could all fall apart, and the world will be much better off.

Bottom Line

McCormick's strongest move is reframing the data center not as a symptom of AI excess, but as the primary vehicle for reindustrializing the physical world, effectively bypassing the need for a traditional government industrial policy. The argument's biggest vulnerability lies in its assumption that private capital will consistently prioritize long-term technological advancement over short-term profit, and that the resulting infrastructure will be adaptable enough to serve needs beyond AI. Readers should watch to see if the "bottomless bid" of data centers actually de-risks hard tech or merely inflates a speculative bubble in the energy sector.

Deep Dives

Explore these related deep dives:

  • The Mother of All Demos

    This 1968 presentation serves as the historical anchor for the article's argument that gaming chips evolved from Apollo-funded integrated circuits to create the massive demand driving today's AI infrastructure.

  • Learning curve

    The article hinges on the economic theory that manufacturing scale drives down costs for technologies like advanced nuclear reactors, a principle explicitly defined by this concept.

  • Strategic Air Command

    This Cold War organization provides the specific historical context for the 'Minuteman' missile systems mentioned, illustrating how defense spending previously funded the semiconductor learning curves that now power data centers.

Sources

Thank God for data centers

by Packy McCormick · Not Boring · Read full article

Welcome to the 2,232 newly Not Boring people who have joined us since our last essay! Join 267,788 smart, curious folks by subscribing here:

Hi friends,

Happy Wednesday!

A couple weeks ago, I asked if you wanted me to start sharing more off-the-cuff notes with not boring world subscribers, and the response was great, so we’re back. It’s a Wednesday afternoon, not my normal send time, but these are meant to be less formal and more, “I noticed something interesting, here are my quick thoughts.” This one happens to be a little longer than it is quick, but it’s one I wanted to get out for two reasons:

People hate AI Data Centers, and I think they’re wrong, even if they don’t like AI.

Because I keep hearing, reading, and seeing that AI Data Centers are funding new technologies before they’ve come down the learning curve, which might be a providentially big boon to Reindustrialization and all of the hard, physical things we want to see in the world.

It’s pretty beautiful that gaming chips that evolved from Apollo-funded integrated circuits are creating a product with so much demand that their houses can pay for all sorts of novel technologies, like the Apollo Program did.

Let’s get to it.

Today’s Not Boring is brought to you by… Deel.

Hiring globally doesn’t have to be complicated

Hiring globally can unlock growth, but local laws, payroll, and compliance often slow startups down.

Deel’s free guide breaks down what an Employer of Record (EOR) is, how startups use EORs to hire internationally without opening entities, and when it makes sense to use one so you can scale with confidence.

Thank God for Data Centers.

There exists a vast pool of technologies that are potentially superior to those we employ today, but which require scale and learning curves to reach their potential.

Advanced nuclear reactors are one such technology - they are more expensive than alternatives today, but manufactured at scale, and benefiting from the learning curves required to get there, may become cheaper than other generation technologies. The cost physics are on their side, and nuclear is reliable, safe, firm, and clean.

The challenge with these technologies, in normal times, is that there is little economic incentive for the buyers who would enable the scale to stick their necks out. Natural gas is cheap and abundant, and it’s not that bad for the ...