NVIDIA is roughly eight percent of the S&P 500 on its own. Broadcom, AMD, Intel, Dell, Seagate, and Eaton are all sitting in there too. Before going looking for a clever way to invest in the physical build-out of artificial intelligence, it is worth knowing how much of it you already hold.
The question came up because Andreessen Horowitz announced a $1.1 billion machine age fund aimed at exactly that build-out: chips, data centers, power, cooling, robotics. Their thesis is that everyone is fixated on the software layer while the real bottleneck sits in the hardware and infrastructure underneath. A listener with a tech-heavy allocation wanted to know whether he is genuinely exposed to that shift, or only to a software layer that could get commoditized as the physical capacity gets built.
Why Hardware Suddenly Feels Safer
The worry about software is reasonable on its face. Barriers to entry are low, competition is fierce, and valuations reflect a lot of optimism. So the instinct to rotate toward something you can touch makes sense. Own the land the data center sits on. Own the company making the cooling systems. Own the physical thing.
That instinct shows up every time people get nervous about valuations. We start asking what the floor is, what this could be sold for if the story stops working. Software has always struggled with that question because there is no hard asset underneath it. Concrete and transformers feel like they answer it.
Tangible Does Not Mean Safe
I would push back on how much comfort that provides. Being physical does not remove the possibility of picking a loser. The infrastructure build-out will produce winners and losers exactly as software does, and a data center with the wrong tenant is not protected by the fact that you could go stand next to it.
Tangibility offers an illusion of certainty. It is not a margin of safety.
Open Your Index Fund and Look
Now to the actual question, which is whether a tech-heavy investor is exposed to this shift. He is, and probably more than he realizes.
The S&P 500 is effectively a technology fund at this point. Look at the large holdings and the AI infrastructure story is already written into them. NVIDIA is designing the chips. Broadcom is in the networking layer. Seagate is storage. Dell is building servers. Eaton is in the power management side of the equation. These are not adjacent bets on the build-out; they are the build-out, and they are already in the portfolio at meaningful weight.
Market Weighting Does the Rotation For You
This is the part I find genuinely elegant about a market-weighted index. As a company grows, it becomes a larger share of the index automatically. NVIDIA did not reach eight percent of the S&P because a committee decided AI was important. It got there because the market repriced it.
So run the listener’s own scenario forward. Suppose software valuations do compress and the profits really do migrate toward physical capacity. Those infrastructure companies earn more, their share prices rise, their market caps grow, and their weight in the index grows with them. If you hold an S&P 500 fund or a total market fund like VTI, your portfolio rebalances toward the winners on its own, without you having to correctly identify which ones they would be in advance. You end up owning whatever turns out to be the latest and greatest, by construction.
What the Venture Money Is Actually Buying
None of this makes the a16z fund a bad idea for the people investing in it. What they are buying is different from what you own. Private capital at this stage is reaching for the smaller, earlier, less proven companies: the firms physically constructing the data centers, installing the cooling, supplying components you have never heard of. That is a wider distribution of outcomes, with bigger winners and considerably more losers, and it comes with lockups and illiquidity on top.
There is also a useful signal in the announcement itself. Serious money committing to infrastructure, which is expensive and hard to unwind once built, suggests some conviction that this build-out is durable rather than a passing enthusiasm. You do not put a billion dollars into things bolted to the ground if you expect to change your mind in eighteen months.
Take the Signal, Skip the Reallocation
So I would read the announcement as confirmation rather than as an instruction. It tells you where thoughtful capital thinks the bottleneck is. It does not tell you that your portfolio has a hole in it.
If you hold a diversified, market-weighted portfolio, you are already positioned for the shift the fund is chasing, at lower cost, with daily liquidity, and without needing to be right about which specific cooling vendor wins. Before making a change here, it is worth pulling up your actual holdings with us, your advisors, and seeing the exposure you have rather than assuming you are missing it. Most of the time, the honest conclusion is that the trade you were about to make is one you already own.
This post is adapted from a recent episode of the Scholar Wealth Podcast. For more perspective on where AI exposure actually lives in a portfolio, listen to the full podcast episode here.