AI Infrastructure Fund: Capturing the AI Buildout Without Picking Winners

An AI infrastructure fund invests in the physical backbone that AI systems need to operate: data centers, semiconductors, networking equipment, power generation, and cooling solutions. The idea is to capture the AI buildout without having to predict which AI software company will win. Alexander Green’s ASI Fund, presented through The Oxford Club, is one specific vehicle that implements this approach.

The thesis is simple. Nobody knows whether OpenAI, Anthropic, Google, Meta, or some startup we have not heard of yet will dominate AI. But all of them need the same physical infrastructure. They all need chips. They all need data centers. They all need power. They all need cooling. A fund that invests in that physical layer captures the AI buildout regardless of which software company wins.

This is the picks-and-shovels approach, named after the idea that the people who sold picks and shovels during a gold rush made more reliable money than the miners. In the AI context, it means investing in the infrastructure rather than the applications.

What an AI Infrastructure Fund Targets

The ASI Fund, as Green describes it, targets several categories of AI infrastructure:

Hyperscale data centers. The Stargate project is deploying $100 billion to build AI data centers across the United States. This is a joint venture between OpenAI, Oracle, and SoftBank. Meta is building a Manhattan-sized data center. Amazon’s Project Blue is under construction. These facilities require massive investment in construction, power systems, cooling, and networking.

Advanced semiconductor fabrication. AI chips are the most complex semiconductors ever manufactured. The fabrication facilities that produce them cost tens of billions of dollars. The companies that supply the equipment, materials, and services for semiconductor manufacturing are infrastructure plays.

High-bandwidth networking. AI systems require thousands of GPUs to communicate simultaneously. The networking equipment that connects them, including switches, routers, and optical components, is a critical infrastructure category. For more on this angle, see our Accelerated AI review, which covers Jason Bodner’s thesis on photonics replacing copper inside data centers.

Power generation. AI data centers consume enormous amounts of electricity. ChatGPT alone uses enough power to run 180,000 American homes every day. Nuclear power, particularly Small Modular Reactors, is emerging as a key solution. Green identifies a “nuclear monopoly” that built 400 mini-reactors for the U.S. Navy. For more on this, see our SMR stocks article.

Cooling solutions. AI chips generate enormous heat. The cooling systems that keep data centers operational are a significant infrastructure category, encompassing liquid cooling, HVAC, and thermal management technologies.

The ASI Fund Specifically

Green’s ASI Fund stands for “Artificial Superintelligence.” It is the specific vehicle he recommends in the presentation. He says you can “get in for less than $15,” which suggests a fund structure with a low minimum investment, possibly an ETF or similar vehicle.

The fund is connected to Executive Order 14330, a real executive order signed in August 2025 that opens 401(k) plans to alternative-asset funds. EO 14330 directs the Department of Labor to clarify that plan fiduciaries may include professionally managed alternative-asset funds in 401(k) lineups. This means that over time, ordinary investors may be able to access the ASI Fund or similar vehicles through their retirement accounts. For more on this, see our Executive Order 14330 article.

For more on the fund itself, see our ASI Fund article.

The Bonus Reports

In addition to the ASI Fund, Green’s subscription includes several reports that identify specific stocks within the AI infrastructure theme:

  • “The AI Superstock” — a company that grew revenue 36,000 percent in three years and partnered with OpenAI, Microsoft, and Meta. This sounds like a semiconductor or AI infrastructure company that has experienced rapid growth from the AI buildout.

  • “The Next Wave of AI” — a company controlling 34 percent of the collaborative robot market with 80,000 systems deployed. This is likely a robotics company benefiting from the automation trend, possibly Teradyne (TER), which owns Universal Robots.

  • “The Nuclear Monopoly Powering AI” — a company dominating Small Modular Reactors, having built 400 mini-reactors for the U.S. Navy. This is likely BWX Technologies (BWXT).

These reports provide individual stock recommendations that complement the broader fund approach. The ASI Fund gives you diversified exposure to AI infrastructure, while the reports identify specific companies with higher-conviction potential.

The Structural Forces

The structural forces supporting the AI infrastructure fund thesis are real and well-funded:

  • The Stargate project: $100 billion deployment
  • Meta’s Manhattan-sized data center
  • Amazon’s Project Blue
  • ChatGPT power consumption: 180,000 homes per day
  • OpenAI’s September IPO at $1 trillion-plus valuation
  • Anthropic’s October IPO approaching $1 trillion

These are verifiable facts, not projections. The AI infrastructure buildout is happening right now, and the capital being deployed is unprecedented.

Considerations

The picks-and-shovels approach is sound, but it is not without risk. The valuations of AI infrastructure stocks have risen significantly, and any slowdown in AI spending could trigger a correction. The fund structure provides diversification, which reduces single-stock risk, but it also means you will not capture the extreme upside of a single winner. The right approach depends on your risk tolerance and investment goals.

For the full analysis of Green’s presentation, see our ASI Fund review. For more on what is included with the subscription, see our Oxford Club article.

If you want to explore the full thesis, you can access the ASI Fund presentation through The Oxford Club.

This is not financial advice. Always do your own research before investing.