A Two-Tiered Market and How It Resolves

Futures are trading close to flat Thursday morning after Wednesday's record close. The Iran news flow is quiet with no new developments since President Trump's pause of Project Freedom on Tuesday evening, but oil is still working lower.

WTI dropped 10.5% on Wednesday to $91.54 and Brent fell 9.8% to $99.12, which were the largest single-session declines for crude in years. Oil is lower again on Thursday morning which is a very promising sign of some resolution of the Iranian situation

The chatter in the financial media is all about indexes at record highs and frothy semiconductor action. That is the obvious story and is fodder for endless speculation about a significant top. The more interesting, and profitable, story is what is happening underneath.

A Two-Tiered Market

The action in this market has split into two clearly different tiers, and the divergence is significant but not well recognized.

The first tier is the Magnificent Seven plus the chip group. Those names have been parabolic. Micron Technology (MU) is up over 100% year to date. Advanced Micro Devices (AMD) is up roughly 80% year to date and added another 15% to 18% on Wednesday's print. Intel (INTC) and SanDisk (SNDK) are at fresh 52-week highs.

The S&P 500 is more than 6% above its 50-day moving average, and the Nasdaq is nearly 10% above its 50-day. By any conventional measure, that group is extended and I’m getting very tired of hearing about how dangerous this is.

The second tier is everything else. The equal-weight S&P 500 (RSP) is just now testing its February highs. The average S&P 500 component has barely moved.

Jason Goepfert's data point Wednesday made this concrete. The S&P 500 closed at a record high for two consecutive days with more than 4% of its components hitting 52-week lows. The only prior occurrence of that in 100 years of data was 1929.

That historical comp is sobering, but the underlying observation is that a meaningful slice of the S&P 500 is struggling and many of the charts have solid support and are not extended.

This split is unusual because of its severity. Typically at a market top, everything goes up together and then everything goes down together. The current setup is closer to a mid-rotation market, where capital is flowing into a narrow group of leaders while a much broader group of names sits in their own private bear market. That is the part of the picture I do not see being addressed.

How This Resolves

The big question is whether this two-tiered setup resolves through a broad selloff or through rotation.

In a broad selloff, the chip leaders correct, and the broader market goes with them. That is the bear case. It is also the historical comparison to 1929 that Goepfert is pointing at. When the leaders break, the rest gets dragged down because there are no offsetting bids and the panic is general.

In the rotation scenario, the chip leaders correct, and the broader market goes the other way. Capital that had been concentrated in the parabolic names looks for the next setup, finds it in the names that have not participated, and the leadership broadens. The headline indexes might be flat or down for a stretch while the average stock plays catch up. That is the bull case.

I do not know which one we will get. But I would point out a few factors that argue for a rotation rather than a broad selloff.

First, the macroeconomic situation is improving, not deteriorating. Oil down 10% in a single day, bonds catching a bid, the Iran situation moving toward resolution, and inflation pressure easing as a result. Those are tailwinds for the broader market, not headwinds.

Second, forward earnings estimates have not started to roll over. There are no EPS estimate cuts on the horizon and no signs that forward estimates are too optimistic. If and when that starts to happen it will require some very aggressive defensive moves.

Why This Capex Cycle Is Different

The third point is one I do not see being addressed in the bubble comparisons. What is really different about the enormous AI capex numbers is that the spending is going toward hard assets such as huge data centers, advanced semiconductors, and all the related materials.

Compare that to the last 20 years and most of the R&D and capex spending has gone into software, into code, into things that do not have a physical footprint. The current cycle is producing a huge supply of physical assets that have a very different lifecycle than intangibles like software.

The Mag 7 is committing to roughly $725 billion in 2026 capex, and the spending is going into data centers, power generation, fiber, advanced packaging facilities, memory fabs, and grid infrastructure. Solaris Energy Infrastructure (SEI) signed a third hyperscaler contract for 600 megawatts last quarter and now has 3.10 gigawatts of secured generation capacity, the equivalent of three nuclear reactors.

Corning (GLW) just announced a tenfold increase in U.S. optical connectivity manufacturing capacity in partnership with Nvidia (NVDA) . Micron MU committed to over $25 billion in capital expenditures for fiscal 2026 with construction-related outlays climbing more than $10 billion the following year.

This is significant for two reasons. The first is that hard asset spending creates a base of tangible economic activity that does not evaporate when there is a change in sentiment. Software spending can be cut to zero overnight. A half-built power plant cannot.

The buildout cycle has its own momentum because the assets need years to come online and the contracts behind them run 10 to 15 years. That gives the broader economy a different kind of support than the late 1990s telecom buildout, which was also hardware-heavy and which arguably did create real long-term value once the dust settled.

The second reason is that the cyclicality of the demand has changed. Memory has historically been a boom-and-bust commodity. The current setup with HBM sold out through 2026 and contract DRAM prices doubling quarter over quarter looks more structural because the buyers are signing multi-year capacity commitments rather than placing spot orders.

The stocks making parabolic moves right now are not riding a sentiment wave. They are pricing in orders for hard assets that are visible for years. This isn’t sentiment running on anticipation of clicks or pageviews like what happened during the internet bubble.

This does not mean the chip stocks cannot correct. They certainly will to some extent but it does suggest that any correction comes through time and consolidation rather than through the kind of crash that follows a software-heavy speculative cycle. The hard asset base puts a floor under the broader economic story even when the stocks themselves take a breather.

This is one of the main reasons I lean toward the rotation outcome rather than the broad selloff. The leadership group has structural underpinnings that the 1929 and 2000 comparisons do not. If they correct, the air comes out gradually, and the capital looking for a home finds the second tier rather than running to cash.

The behavior to watch for is what happens when the chip group has its first real down day. If the rest of the market holds up or rallies, the rotation is happening. If the rest of the market sells off in sympathy, the broad selloff scenario is in play. We have not had that test yet. The chip group has not had a down day to speak of in two weeks.

The RSP Chart Is the Tell

The single most useful chart for navigating this is the equal-weight S&P 500 (RSP) versus the S&P 500 (SPY) .

RSP has lagged for months and is just now back at its February highs. If RSP breaks out to fresh highs while SPY churns or pulls back, the rotation thesis is confirmed and the leadership is broadening. If RSP fails at the February highs and rolls over while SPY also weakens, the broad selloff thesis is winning and the action gets ugly.

I have not seen this divergence addressed much. Most of the commentary is focused on whether the index is at a top, which is the wrong question. The right question is what happens to the average stock when the leaders take a break, and the RSP chart is going to answer that question in real time.

My Game Plan

The plan is to keep working the rotation rather than guess at the indexes. The chip leaders are extended and not where I want to chase. I’m focusing instead on the secondary names and small-cap earnings in particular.

Watch the equal-weight S&P 500 as a measure of rotational action and also watch sectors like biotechnology, finance, retail, energy, etc. to see where cash is flowing. The setups in stocks that pulled back into support and are starting to firm up are the higher quality opportunities right now.

The bears are convinced that a correction in the most extended stocks is going to destroy the entire market. The bulls believe that the market can handle the pockets of overbought action with rotational action. Either way, the answer comes from the charts, not from the commentary. Watch the divergence between RSP and SPY. That is the question this market needs to answer.

At the time of publication, Rev Shark had no positions in any securities mentioned.