The Proper Way To Play AI

The buildout looks familiar, the stakes look bigger, and the state is moving closer to the table. Here is how to think about owning it.

A historical comparison

In 2000, Cisco Systems briefly became the most valuable company in the world. They were a roughly $550 billion symbol of an internet buildout racing ahead of its own demand. Today's AI infrastructure boom invites the same comparison, and its cheerleaders embrace the scale. Nvidia CEO Jensen Huang has called it "the largest infrastructure buildout in human history." The internet did eventually deliver everything it promised, but it took years longer than the market priced in, and many who funded the first wave never lived to see it. The lesson is not that the technology fails, but rather that it is that timing and survival decide who gets paid.

The contrast with that era is real, and worth weighing:

• Then: debt-funded carriers laid fiber that sat largely dark, with utilization reportedly below three percent.
• Now: the spending is largely financed from the cash flows of the biggest companies on earth, with hundreds of millions of users already in place. The catch is size, since annual capex now dwarfs the entire telecom buildout, and chips age far faster than glass.

State intervention

The second force is the state. Washington already holds a roughly 10 percent stake in Intel, and reporting this year says OpenAI discussed offering the government a five percent stake of its own. The presiding President, Donald J. Trump, has said no to nationalization but "I might" to an equity stake. The thesis for deeper entanglement is easy to state: these firms hold too much value to be allowed to fail. They sit at the center of surveillance capability, geopolitical competition with rival nations, and the productivity gains that could define the next technological edge.

The counterargument is that officials, including the White House AI czar, have said there will be no federal bailout and that other firms would simply take a failed one's place.

Investors should hold both possibilities in mind, because a backstop would cushion a downturn but also reshape who captures the upside.

The play

Which brings us to the proper way to play it. Unlike farmland or wheat, AI is a wild vehicle, one that delivers excitement and shocks in equal measure. It may favorably merit a place at the table, on disciplined terms:

1. Size it as a satellite, not the core. Keep stable, hard assets as ballast and let AI be the volatile sleeve.
2. Favor cash flow over narrative. The survivors of the last cycle were the ones who could fund themselves through the shakeout.
3. Build the position in stages. Buying over time softens the cost of mistiming a market that can run in either direction.
4. Respect the political layer. Government involvement is now a variable in every valuation.