A market that punishes both spending and restraint calls for a strategy that can survive the market games. SPY is well positioned here at this point.

There is a certain kind of market that makes investors feel foolish regardless what they do. If you've been actively trading or trying to position in past few weeks, you know what I mean.

If a company spends too aggressively on AI and investors worry about financing. If it slows spending and investors worry about growth. Earnings beat expectations, but the stock falls because cash conversion disappoints. Bond yields decline, providing some relief, until someone asks whether they are falling because the economy is weakening. It's a painful, lose-lose situation.

It feels like heads you lose, tails they win. The explanation arrives after the move, and the next explanation can contradict the last one.

The recent ZeroHedge discussion of Goldman's trading-floor commentary captures that frustration. Its useful observation is that AI demand can remain strong while the economics of financing the buildout deteriorate.

My concern is how to react and wanted to share my perspective. A persuasive explanation for a selloff can quickly become a reason to abandon every position associated with it and I will spare you from the lamest punt of all - "cash is always a position". It's not that its invalid, its just that it doesn't do much when you are in positions or want to be exposed to markets with some degree of risk.

Before constructing a portfolio around the next disaster, there is a simpler question worth taking seriously: what if owning SPY already addresses a meaningful part of the problem?

Why good demand can produce disappointing returns

The AI financing dilemma is real. Computing capacity must be paid for before much of the revenue arrives. Servers depreciate, power costs money, and customers do not necessarily pay on the same schedule that lenders expect repayment.

Higher financing costs change the return available to shareholders even if customer demand remains intact. A project can still be worth building while becoming less attractive for the people who originally supplied its equity.

Consider a deliberately simplified business with $100 of enterprise value, $60 of debt and $40 of equity. If its enterprise value falls 15% while debt stays unchanged, equity falls to $25. The business lost 15% of its value; shareholders lost 37.5%. No collapse in demand was required. (Watch IWN for debt perception on equities as small caps reflect this.)

This is also why the distinction between a profitable supplier and a heavily financed infrastructure owner matters. Both may be labeled “AI,” but their shareholders own very different claims on the same spending cycle.

There are limits to the bearish argument, too. Future lease and purchase commitments are not all debt due tomorrow. Customer prepayments can reduce funding risk. Whether those arrangements are healthy depends on contract terms, counterparties and project economics.

The productive question is how much cash reaches each share after the buildout is funded. A large backlog does not answer it.

The case for owning both sides of the rotation with SPY

SPY gives investors exposure to the companies building AI infrastructure and to many businesses that might benefit from using it.

If AI investment generates attractive returns, the largest technology holdings can continue contributing to index earnings. If investment becomes more disciplined and inflation pressures ease, lower yields can support businesses elsewhere in the economy. If computing gets cheaper, some of the value may migrate from infrastructure providers to customers.

These outcomes are not guaranteed, and cheaper AI will also disrupt some existing businesses. But they create several bullish vectors for a portfolio.

This is the strongest argument for SPY in a apparant "rudderless market": an investor does not have to identify every leadership change before it happens.

There is an important mechanical element to SPY. SPY does not actively sell expensive winners and buy cheap laggards. Its market-cap weights largely move with prices. It benefits from owning the eventual winners already rather than from a manager making every rotation call correctly.

As of this week, the Mag 7 represented approximately 34% of SPY with the reaming 493(!) companies represented approximately 66%. That is meaningful diversification, even though broader AI exposure extends into semiconductors, power equipment and other industries.

With this in mind, here is a model with some some likely outcomes. (Personally, I think the outcome is somewhere between broadening and AI resumption which perhaps is the reason for this blog to begin with.)

And given these AI-derived scenarios, you can see that SPY holds strong in 3 of the 4. The appealing version of the rotation notion is if AI investment cools, financing pressure eases, bond yields decline, and the rest of the market gains room to grow.

The painful outcome starts with canceled projects, layoffs, weaker consumption and credit losses. Rates may decline in that environment, too, but earnings can fall faster than financing conditions improve.

SPY can spread exposure across changing winners. It cannot diversify away an economy-wide earnings decline.

How I would position?

I've already mentioned this above a bit, but to clarify/harden.

SPY would provide the core equity exposure. It allows participation if AI leadership resumes and if earnings growth broadens. Moving from SPY into consumer discretionary or other boring "safe stocks" isn't going to give you the upside juice, but only the downside.

Treasuries could be ok, if just want to wait things out and gives capacity to buy when prices improve and you have to time the market if you think things are that bad. Waiting has a cost if markets rise, so cash needs a purpose beyond an indefinite expectation of a better entry. This just doesn't seem to be "the way" for Mometic readers.

Individual stocks would have to earn their place. There are some AI plays that have been crushed, but debt concerns and AI overhang may be tough to time. (IREN and ilk). A stock falling 30% (or 65% in IREN) does not establish that it is inexpensive given the bond loss scenario mentioned above. SpaceX (SPCX) seems to be interesting for potential momentum. There is another 300m shares being released to market that will likely pressure shares for a few days or more, and the TSLA merger weighing, but many of the numbers impressed.

Position size should follow the consequences of being wrong. A 5% position losing 40% costs approximately 2% of the starting portfolio. At a 20% allocation, the same mistake costs 8%, before any correlated losses elsewhere.

For new equity purchases, I would establish exposure in stages. Later purchases could follow either an improved valuation with intact fundamentals or stronger evidence of cash generation. A lower price accompanied by worse economics is a reason to recalculate, not an automatic instruction to average down.

A one year view

Here is probability of likely outcomes over next 1 year.

Multiplying each assumed return by its probability produces an expected return of approximately 4.1%. These are AI-based models and at this point I think as good as anything.

The most revealing part is if the AI bust remains contained and SPY finishes flat in that scenario instead of falling 15%, expected return rises to approximately 7.8%. Nothing else changes. In other words, the money has to go somewhere and if AI "fails" per se, the market will dive into all those positions which were left for dead from the AI revolution. Can the rest of the economy absorb the slowdown, or does financing stress become a broader earnings problem?

What would change my position?

I would watch whether funding becomes more available, whether backlog converts into collected cash, and whether earnings estimates outside the AI world improve. I think companies like CAT, DE, JNJ, RGEN, AMZN (retail side), and HD. Market breadth adds insight. A rising equal-weight index (RSP) alongside stable earnings expectations would support the rotation thesis.

For active traders, there may be opportunities in each shift. I'm not sure which bucket of stocks I would pick for trying to time each swing, but its doable.

For investment research, the standard is equally demanding: a selected company should offer something better than the broad market after accounting for downside, concentration and the work required to hold it. That is a useful discipline for any research process, including ours.

“Just own SPY” sounds almost offensively simple given our product line and in time will become invalid. But the current market does not really justify picking individual stocks, just yet. Perhaps mid terms need better clarity first.

Also - SPY is only suggested for a chunk of your portfolio - a buoy if you will. The goal is to remain invested through changes in leadership without depending on every market explanation being right. With our current administration, the wars, elections, and AI overhang - I think its a good of a hypothesis as any.

Brent @ Mometic.com