Is Your Retirement Exposed to the Biggest Speculative Bet of All Time?

If your retirement savings are mostly in U.S. stocks indexes, you are party to the same extraordinary and historic wager the CEOs of Alphabet and Amazon are making. The difference is, for them there is plenty of upside and no meaningful financial downside given their wealth and status. For retirement investors, much of the upside may already have been captured in stock prices and if AI spending doesn’t pan out, they could suffer financially. That’s worth digging into.

Four companies — Microsoft, Alphabet, Amazon, and Meta — are on track to spend roughly $750 billion this year building artificial intelligence infrastructure. For perspective, the entire budget for the government of France is about $900 billion and for Canada it is about $425 billion. We are seeing the largest concentrated capital investment in the history of private enterprise. These are bold historic bets.

Last week I made the case that the market’s check engine light is on. This is a look under the hood at what has the engine running so hot.

The vast majority of retirement savings are invested in some version of passive index funds. The optimism around the AI bet has contributed to the present condition: the ten largest companies now make up nearly 40% of the S&P 500 — the other 490 companies share the remaining 60%. If you own an S&P 500 index fund or similar strategy, you own a concentrated bet on a handful of companies, and those companies have concentrated on one idea.

Few doubt AI will be transformative. What is not understood — what nobody can tell you — is whether these specific companies will earn a good return on three-quarters of a trillion dollars or if developments will turn painfully against them. They are making the wager, but you don’t have to. 

The "too hard" pile

Warren Buffett says he has three boxes on his desk: In, Out, and Too Hard. Most of what crossed his desk went in the third box. Not because he wasn't smart enough to analyze it — because he'd learned that the intelligent move, when the outcome of something genuinely cannot be predicted, is to decline to predict it.

Whether these four companies earn back $750 billion, on what timeline, and who ends up capturing the value — that is a "too hard" question. This week in New York I spoke with a portfolio manager for the largest US growth fund about these very questions. He knows these companies and the CEOs very well, yet he echoed Buffett’s advice—saying “humility is wisdom.” His portfolio is not concentrated like the indexes.

Why spend so much?

Let’s understand what is behind the bet these CEOs are making. If you use AI today, you probably use it the way most people do: to draft an email, summarize a document, answer a question. It's useful. It's not $750-billion useful. That kind of task doesn't require the largest buildout in history.

So the spending only makes sense if they're building for something you're not doing yet. They are. The bet is that AI stops merely answering and starts doing — running a task from start to finish, on its own, for as long as it takes. The industry calls this "agentic" AI, and it is far more useful, and far more compute-hungry, than the chatbot you know.

For example, every business has ideas about what would be an amazing service or feature to offer but is not worth paying a person to do. AI might do those things and massively increase revenue and profits. Maybe...

A specialty food company sells four thousand gift boxes a year. Every buyer is a likely repeat customer — for the next holiday, the next anniversary, the next occasion — but no business can afford a salesperson to track one $90 customer. An AI agent might do it. Imagine that the “agent” notices the one-time customer posted about his wife’s girl’s trip to Napa, writes an email to suggest sending a special food basket and asks if it can help you track down the hotel address. Multiply that across every business in the country.

H-E-B knows you buy roughly the same things every Saturday. Today that knowledge sits idle. An AI agent could assemble your regular list Friday night, ask you to approve it with one tap, suggest a delivery window that fits your calendar — and mention that since you like sparkling water, there's a new Waterloo flavor you'd probably enjoy.

Over at the H-E-B pharmacy they know exactly which patients stopped refilling their prescriptions. Today the frazzled pharmacist has no time to call and investigate. With agentic AI, a solution is possible and inexpensive to implement.

AI can be cheap, reliable, and tireless — doing endless small, valuable tasks that were never economical to hand a human. That is a big part of the vision. If it arrives, $750 billion will look modest.

What could go wrong?

The vision might not play out as described for any number of reasons. Here are a just two:

The killer apps might not show up, or might take far longer than expected. Today's AI agents work impressively for a while and then make an avoidable mistake, and nobody hands their customer relationships to something that's reliable most of the time. Closing that gap is the whole bet, and it isn't closed. Meanwhile the bill for the infrastructure comes due on a fixed schedule, whether or not the revenue shows up on time.

Or the technology gets dramatically cheaper — which sounds good, but isn't, if you own the builders. If a breakthrough sharply cuts the computing power AI needs, or if free, openly available models become nearly as good as the expensive ones, the advantage shifts away from whoever spent the most. The buildout still gets used — but it earns ordinary, utility-like returns.

What to do when you don't know

The CEO’s will be judged by the payoff related to this historic AI buildout. Your mission is to manage your retirement. If you see the exposure you have in your holdings and wonder if you are betting correctly, you are playing the wrong game. Instead of worrying or trying to predict, prepare with the following rational responses:

Keep near-term money out of risk assets. Anything you'll need to spend in the next two years shouldn't ride on how a technology transition plays out. That should remove the most dangerous outcomes.

Don't let your portfolio become a concentrated bet on one story unfolding perfectly. It might. It probably won't unfold cleanly. The trouble is that a standard index fund has become that concentrated bet — you didn't choose it; the math of a rising market chose it for you. Nobody is forcing you to own the index exactly as it's built. That's a choice you're allowed to revisit.

Own the beneficiaries, not just the spenders. The portfolio manager I mentioned is a big believer in agentic AI, but one of his highest conviction stock picks right now is not a hyperscaler, it is a very old consumer health care brand he thinks is poised to benefit from AI advancements. He owns some but not all of the companies making existential bets. As a seasoned world class investor, he sees that if agentic AI delivers even part of what's promised, a far wider set of businesses will benefit— companies outside of tech that get more productive and more profitable using these tools without betting their survival to build them. They exist in public and private markets. There are real estate themes tied to this too — the power, the buildings, the physical footprint of all this computing has to live somewhere. In short, get truly diversified.

If you were the CEO of Google, this massive bet on AI is completely rational with huge upside and little personal risk. If you are trying to protect your retirement, I would suggest your risk and reward of predicting the outcome is quite different—so don’t predict, prepare.

If you have questions about today's content, please write us at info@networthypublishing.com.

Please note, this material is provided as general educational content, not personalized investment advice; your own situation deserves its own analysis. Nothing in this article should be considered specific investment advice.