James
Notaris

The popular account of the artificial-intelligence boom is a story about chips. Faster processors, bigger clusters, the race for the most advanced silicon. It is a compelling narrative, and it is incomplete. The binding constraint on how fast this industry can grow is not the chip. It is the electricity to run it and the grid connection to deliver that electricity to the building.

A single large AI training cluster can draw as much power as a small city. Multiply that across the data centers being planned worldwide and the demand curve bends in a way the grid was never built to absorb. You can buy all the accelerators you want; if you cannot energize them, they are inventory, not capacity.

Compute is abundant in theory and scarce in practice, and the thing that makes it scarce is power.

This reorders where value sits. The most valuable asset in the chain is no longer purely the technology. It is the large, energy-rich site with real grid capacity, secured interconnection, and the ability to deliver reliable power at a competitive price. Those sites are finite. Permitting takes years. Interconnection queues stretch out further. Transmission cannot be conjured on a quarterly timeline. Scarcity of this kind does not resolve quickly, and scarcity that does not resolve quickly is exactly where durable value accumulates.

The lesson of electrification

There is a precedent worth sitting with. The defining inventions of the twentieth century did not arrive the moment they were imagined. They arrived once the power to run them had been built. The assembly line, the elevator, the refrigerator, the computer itself: each waited on an electrical grid capable of carrying it. The breakthrough was downstream of the infrastructure, not the other way around.

Artificial intelligence is no different. The models are remarkable, but they are tenants in a physical world that has to be wired first. Whoever controls dependable, low-cost power controls the pace of the entire revolution, because everyone downstream is waiting on the current flowing into the building.

That is why capital is moving toward power the way it once moved toward oil: behind-the-meter generation, repurposed industrial sites with existing electrical infrastructure, long-term energy contracts, and the unglamorous work of interconnection. It is less visible than the latest model release, and considerably more decisive.

The companies that understand this are not asking how to buy the most chips. They are asking how to secure the most power. In the AI era, that is the question that determines everything else.

There are two ways to participate in the artificial-intelligence buildout. You can operate the compute, buying accelerators and renting them out as cloud capacity, or you can own the physical foundation those operators depend on and lease it to them. The two look similar from a distance. Up close, they are different businesses with different risk profiles and very different durability.

Operating compute is a hard place to stand. The hardware depreciates quickly, the next generation is always months away, and price competition is relentless. You are exposed to technology risk at the most volatile layer of the stack, in a market where today's premium product is tomorrow's discount.

Owning the durable layer

The landlord owns something slower-moving and more defensible: the power, the cooling, and the buildings. That capacity is leased under long-term contracts to the firms racing to train and run the next generation of models. The economics look less like software and more like industrial real estate, with recurring, contracted cash flows tied to something every AI company needs and almost none can build for itself.

The operators fight over a depreciating asset. The owner of the building collects rent through the entire cycle.

History rewards this position with unusual consistency. In the gold rush, the durable fortunes were made selling tools and provisions, not panning for gold. In the early internet, the firms that laid fiber and built the data centers quietly underwrote everything that came after, long after individual applications rose and fell. The pattern repeats because it reflects something structural: infrastructure outlives the products that run on it.

This is not an argument that operating AI compute cannot be lucrative. It is an argument about where risk is concentrated and where value is durable. The landlord converts a volatile technology race into stable, contracted income, and sits beneath the entire industry rather than competing on its most exposed front line.

As the AI buildout matures, the question of who captures the lasting value will not be settled by who had the best model in any given year. It will be settled, in large part, by who owned the ground it ran on.

To understand Bitcoin, you have to start with 2008. Not as background, but as cause. The two are inseparable, and anyone who tries to explain one without the other is telling half the story.

The financial crisis was not, at its root, a failure of mathematics or even of regulation. It was a failure of trust. Trust in the intermediaries who were supposed to manage risk and instead concentrated it. Trust in instruments no one fully understood. Trust in the idea that the institutions at the center of the system would behave as their incentives quietly told them not to. When that trust broke, it broke fast, and the cost was socialized across people who had no part in creating it.

2008 was the breakdown of trust in centralized finance. Bitcoin was the attempt to build a system that needed less of it.

A few months later, in January 2009, the first block of the Bitcoin blockchain was mined. Embedded in it was a newspaper headline about bank bailouts. That was not decoration. It was a thesis. The proposition was simple and radical: a financial system whose integrity did not depend on trusting any single institution to behave well, enforced by rules and mathematics rather than by reputation and regulation.

The arc since

For years, the serious people dismissed it. It was a curiosity, a speculation, a thing that would not last. Then the arc bent. The same institutions that had waved it away began to hold it. Spot exchange-traded products brought it onto the most conventional balance sheets in the world. The asset that was born as a protest against the financial establishment became, in time, a holding of that establishment.

There is an irony in that, but also a lesson. The crisis exposed a problem, and the response was an attempt to engineer the problem away. Whether one views Bitcoin as money, as an asset, or as an experiment still running, its origin is not in doubt. It was a direct answer to a specific failure, written by people who had watched that failure up close.

Having lived through both the collapse and the rise that followed, I find the throughline impossible to ignore. The story of the last seventeen years is, in large part, the story of what happens when trust in a system fails and someone decides to build one that asks for less of it.

Every fan knows the box score. Almost none of them know the tax return — and in professional sports, the tax return is its own brutal competition.

Start with baseball. A major-league season runs 162 games, roughly half of them on the road, scattered across the better part of two dozen states and cities. Here is the part that surprises people: an athlete owes income tax in nearly every place he earns money, which means nearly every place he plays. It is called the jock tax, and it turns a ballplayer into one of the most complicated taxpayers in the country. The income is sliced up by “duty days” — how many of the season’s working days were spent in each jurisdiction — and returns are filed in all of them. A single player can file a dozen or more state returns in one year, plus city taxes in the places that levy their own.

The doctrine has a famous origin. After the Chicago Bulls won the 1991 title, California taxed the visiting players on what they earned in the state. Illinois answered with a reciprocal levy the press cheerfully nicknamed “Michael Jordan’s Revenge.” What started as a feud between two states became standard practice everywhere, and today it quietly shapes the take-home pay of every athlete who travels for a living.

Tennis plays a harder game

If baseball is a fifty-state puzzle, tennis is a global one. A touring professional is not a team employee; she is an independent contractor running a worldwide business, competing in a dozen countries a year and earning prize money, appearance fees, and endorsements in each. That means foreign withholding, tax treaties, and the genuinely thorny question of how much of a global endorsement contract any single country may tax.

When the tax code can keep a champion off the court, it is no longer a footnote. It is part of the game.

Some nations reach further than others. A handful tax not only the prize money won inside their borders but a proportional slice of a player’s worldwide endorsement income, apportioned by how many days the athlete trained or competed there. The consequence is real and public: some of the world’s best players have openly limited their appearances in certain countries because the tax on showing up exceeded the prize for winning.

What it reveals

Both sports expose the same truth: our tax system was built for people who stay put, and it strains when applied to high earners who move for a living. The athlete is just the most visible version of a much larger story — the taxation of mobility — that now reaches remote workers, founders, and anyone whose income follows them across borders.

It is also, frankly, a thrilling problem to solve. The athletes get the highlight reels. The people who keep their finances intact across fifteen jurisdictions never make the broadcast — but they are playing a championship game of their own.