Investment Gems

I am getting more uncomfortable with the macro setup.

For a while, the story was actually pretty constructive. Energy scarcity was pushing up commodity prices, but higher prices were also improving the expected returns on new investment. Machinery was working. AI infrastructure was working. Financials were holding up despite higher long rates. It looked like we were moving from a simple scarcity shock into something more durable: scarcity creating profits, profits creating capex, capex creating financing demand, and the whole thing turning into a broader capital-formation cycle.

That is still possible. But the latest tape is beginning to look different.

The center of gravity has moved toward refining, shipping and physical throughput. Brent is above $100, refined-product scarcity is becoming more acute, Hormuz traffic remains badly impaired, and the strongest equity leadership is increasingly concentrated in the companies that benefit directly from shortages. At the same time, machinery, financials and parts of the AI hardware complex are no longer absorbing higher oil and higher yields quite as easily as they were a few weeks ago.

Regime signal	Earlier phase	Current phase	Read
Crude / energy	Strong	Very strong	Scarcity intensifying
Refiners	Strong	Extremely strong	Throughput bottleneck
Tankers / logistics	Strong	Strong	Physical disruption
Oil services	Improving	Lagging refiners	Supply response not yet dominant
Machinery	Leading	Losing relative strength	Warning
AI hardware	Resilient	More selective / softer	Warning, not break
Financials	Holding high yields	Weakening	Important warning
10Y yield	High but tolerated	Near 5% and restrictive	“Good yields” becoming “bad yields”
Gold	Strong, then mixed	Rate-sensitive	Not yet full monetary-panic regime

The important thing is that this is no longer just an oil story. If the only thing moving was Brent, I would be much less worried. The equity tape is telling us that the real bottleneck is downstream. Refiners are ripping. Tankers are working. Product markets are tight. Insurance and shipping costs through Hormuz remain extreme. Fertilizer, chemicals and other parts of the physical economy are also showing up in momentum screens.

That suggests the market is pricing a shortage of conversion capacity and physical throughput, not merely a temporary shortage of crude. And that is potentially much more economically damaging.

The economy does not consume Brent futures. It consumes diesel, gasoline, jet fuel, chemicals, fertilizer, transportation and electricity. That is why I have been bullish on Delek Holdings. If those bottlenecks remain tight, the inflationary impulse reaches businesses and households even if crude supply itself is not catastrophically impaired.

That is where Scarcity Reflation can flip from constructive to destructive.

The bullish version of the thesis is straightforward. High commodity prices improve returns on new capacity. Producers respond. Services companies start getting orders. Machinery demand rises. Infrastructure gets built. Banks and capital markets firms benefit from financing the buildout. Eventually new supply arrives and the economy absorbs the shock through investment rather than recession. That would bring adjacent energy companies along. I have hinted at that for Denison Mines and I still believe it to be true, if we are in a capital formation cycle there are plenty of opportunities for really great asset.

The problem is that the market is not yet giving us enough evidence that this transition is happening.

If that constructive sequence were taking hold, I would expect oil services such as SLB, FTI, AESI and NESR to increasingly outperform the refiners. That would mean producers are saying, in effect: these prices are high enough, let’s drill, build and expand.

So far, the opposite signal remains stronger. Refiners are still leading services. That tells me the market is rewarding scarcity rents more than scarcity relief.

The second confirmation I want is machinery. If this is genuinely becoming a capital formation cycle, then CAT, DE, CNH, AGCO, ETN and VRT should eventually reassert leadership. Those are the companies that benefit when the response to scarcity becomes physical investment.

They matter because they are one step removed from the commodity itself. A refiner can rally because the shortage is getting worse. Caterpillar should rally when someone decides to do something about it. That is a very thing.

The same logic applies to AI. I still think AI industrialization is one of the strongest secular capex themes in the market. But the macro question is whether it can remain strong when oil is above $100 and the 10-year is near 5%.

Earlier, the answer was clearly yes. AI infrastructure was acting almost like a separate industrial cycle. Memory, servers, networking and semiconductors could continue higher even while the cost of capital rose. Now that resilience is being tested.

One weak session in NVDA or MU does not invalidate the thesis. But if AI hardware starts rolling over together with machinery and financials, then the macro constraint is becoming stronger than the secular capex story. That would be a major problem.

The financials are probably the most important signal of all.

For months, the rise in long yields could be interpreted positively. If JPM, IBKR, BGC, PJT and XLF were healthy while the 10-year moved toward 5%, that suggested the economy was simply demanding more capital. Even financial services like PaySign were leading with huge strenght. Rates were high because nominal activity, investment and financing demand were high.

Those were “good” high yields.

But if the 10-year sits at 5% while financials weaken, credit spreads rise and capital-markets activity deteriorates, then the same yield means something completely different. Those become “bad” high yields. The cost of capital is no longer reflecting economic strength. It is beginning to choke it. That is the macro fork we are approaching.

Don’t ask whether rates are high. Ask what can coexist with high rates.

If the 10-year is around 5% and machinery, AI infrastructure and financials are making new highs, we are probably in a capital boom. If the 10-year is around 5% and the only things making highs are refiners, tankers, energy producers and gold, then we are probably in an inflationary scarcity shock.

Right now, the marginal tape looks more like the second description.

That does not mean the whole Scarcity Reflation thesis was wrong. In fact, the first half of it has been extraordinarily right. Scarce physical assets are being repriced aggressively. Refining capacity is valuable. Shipping capacity is valuable. Energy infrastructure is valuable. Hard assets are valuable.

The question is whether scarcity is now becoming too successful. There is a level where high prices stop encouraging investment and simply destroy affordability. There is also a level where high yields stop reflecting strong capital demand and start killing capital formation.

We may be getting close to both. That makes the next rotation extremely important. If leadership moves from refiners toward SLB/FTI/NESR, then into CAT/DE/ETN/VRT, and finally back into JPM/BGC/IBKR/PJT, I would become much more constructive again. That would mean the market is progressing through the full chain:

scarcity → profits → investment → financing → new capacity.

But if leadership remains trapped in refiners, tankers and energy while machinery, AI hardware and financials continue deteriorating, then the sequence becomes:

scarcity → inflation → higher yields → demand destruction.

That is a much uglier regime. From an investment perspective, I would therefore avoid treating every “scarcity” asset as equally attractive. Refiners and physical-throughput beneficiaries still deserve respect because the tape is confirming the shortage. Selective midstream and tanker exposure also makes sense while Hormuz remains impaired.

But I would be careful about simply adding more broad cyclicality on the assumption that expensive oil automatically creates an investment boom. It only does if the rest of the economy can finance and absorb that investment.

For now, I would keep the highest-quality AI infrastructure names, because the secular buildout remains intact, but I would stop pretending they are immune to macro. I would keep machinery on watch rather than aggressively add. And I would pay extremely close attention to financials, because they may be the earliest signal that 5% long rates have crossed from productive to restrictive.

The macro thesis has therefore evolved.

A few weeks ago, I was asking whether Scarcity Reflation could become a durable capital-formation cycle, but today, I think the more urgent question is whether the scarcity shock is beginning to eat the capital cycle before it can fully form.