The Monday State of Play: The Market Can’t Pick a Story
AI spending is accelerating, semiconductor stocks are cracking, and oil still doesn’t quite believe the war.
Happy Monday, everyone. Hope you all had a fantastic weekend and caught at least some of the World Cup final between Spain and Argentina, because that one got a little spicy.
And speaking of spicy, the market is being asked to hold two fairly uncomfortable ideas at the same time.
The first is that the AI infrastructure buildout continues to accelerate, with Microsoft becoming the latest major technology company to adopt AMD’s Helios rack-scale system, even as semiconductor stocks have fallen into a bear market and investors are suddenly being encouraged to look somewhere other than AI hardware.
The second is that the Middle East conflict escalated sharply again over the weekend, the ceasefire the market had been leaning on has largely unraveled, and Brent briefly pushed back above $90. This presented a reminder that the geopolitical premium traders kept giving back on every diplomatic feeler was never quite as safe to sell as they had decided it was.
Both stories can be true for a while, but probably not forever.
The AI trade isn’t dying, although the days when every dollar of hyperscaler capital spending automatically translated into a higher multiple for every company attached to a server rack may be behind us. Investors are beginning to distinguish between the companies building the infrastructure and the companies that can actually use it to generate revenue.
It doesn’t help that the latest leg down had a specific trigger. Moonshot’s new Kimi K3 model revived the oldest fear in this trade: that competitive AI capability may arrive far more cheaply than the hyperscalers are spending to build it. It’s another DeepSeek echo, and it puts the burden of proof squarely on revenue, productivity and inference costs rather than raw capacity.
That makes Alphabet’s earnings Wednesday the most important event of the week.
We already know Google is spending heavily. The question is whether inference demand, cloud adoption and AI-driven products are beginning to produce enough revenue to justify the capital being deployed across the industry. If Google Cloud accelerates and management raises spending expectations, the market will have an easier time believing the AI buildout is broadening rather than peaking. If revenue disappoints while spending remains elevated, investors may begin asking whether the industry is building capacity faster than customers are willing to pay for it. Alphabet reports Wednesday, July 22.
Oil is dealing with a different version of the same problem, and this weekend it became a lot less abstract.
Brent briefly moved above $90 as the United States entered a ninth consecutive night of strikes, traffic through the Strait of Hormuz slowed to just four vessels on Sunday and the Houthis declared a naval blockade against Saudi Arabia. Iran acknowledged receiving another ceasefire proposal, which was enough to pull crude back below $88, but the fighting and shipping disruption continued.
For weeks, the market treated every hint of negotiations as a few dollars off the price of crude. It is now being forced to test that assumption against what is actually happening on the water.
I don’t think the move above $90 means traders suddenly believe the worst-case scenario. It means the assumption that the disruption would remain temporary is finally being challenged. Geopolitical fear can disappear quickly, while physical shortages are considerably more stubborn, and if Hormuz traffic remains impaired and refined-product inventories begin tightening, the physical market will eventually force the paper market to pay attention.
That’s the part I’m watching—not the daily diplomatic headlines, but the shipping volumes, tanker activity and diesel inventories that tell us whether this remains a scare or becomes a shortage.
AMD’s Helios announcement reinforces the idea that the next phase of the AI buildout may reward a broader group of suppliers, although it doesn’t suddenly erase Nvidia’s lead. At the same time, the rotation away from AI hardware has become a very popular story, which is usually when we should stop congratulating ourselves for discovering it. The semiconductor index is already more than 20% below its recent high, so the easy version of that trade has probably already happened.
Domino’s offered a smaller but useful read on the consumer Monday morning. Revenue held up, but U.S. same-store sales rose just 0.1% as customers continued hunting for value. That isn’t a roaring consumer reacceleration, but it does suggest households are still willing to spend when the price and promotion make sense. The consumer is alive, but still likes a coupon.
Global Macro
Oil remains the biggest variable. Hormuz traffic has slowed sharply, the conflict has broadened and the market is still hoping diplomacy arrives before inventories become a problem. I’m less interested in every proposed ceasefire than I am in whether shipping volumes and refined-product inventories begin confirming a more persistent disruption.
The industrial cycle still needs confirmation. Expectations for a broader cyclical recovery remain alive, but the evidence is mixed. Texas Instruments, Freeport and United Rentals should give us a much better idea of whether demand is actually improving beneath the narrative.
Liquidity remains the caveat. The backdrop still leans reflationary, but conditions haven’t improved enough to make every cyclical setup attractive. We can remain positioned without chasing every move simply because a new narrative appeared on television.
Earnings This Week
Alphabet — Wednesday: This is the week’s main event. Cloud growth, AI monetization and capital-spending guidance should tell us whether the revenue side of the AI story is beginning to catch up with the infrastructure side.
Texas Instruments — Wednesday: TI remains one of the better reads on industrial and automotive demand. Orders, inventories and management’s outlook should tell us whether the cyclical recovery is becoming visible in actual activity.
GE Vernova — Wednesday: The EPS number matters less than orders, backlog and delivery timelines. Power and grid investment remain long-duration themes, but the stocks have had enormous runs and now need execution to do more of the work.
Freeport-McMoRan — Thursday: Freeport remains one of the cleaner reads on copper and physical demand from China. The structural case for copper is easy to make; the harder question is whether current demand is strong enough to support today’s expectations.
RTX — Thursday: With the United States actively engaged in the Middle East, backlog growth and production capacity should provide a useful look at how quickly defense demand is translating into actual deliveries.
So Where Does That Leave Us?
Okay, so the market isn’t giving us one clean story this week, and I’m not sure we need to force one.
AI spending continues to accelerate, but the hardware trade is being repriced as investors demand evidence that all of that infrastructure will eventually generate an acceptable return. The Middle East has escalated again, and oil (after weeks of assuming diplomacy would arrive before inventories became a problem) is finally being forced to test that assumption. The cyclical recovery remains plausible, but this week’s earnings need to show that it exists somewhere beyond the forecasts.
Alphabet should tell us whether AI monetization is beginning to catch up with AI spending. Texas Instruments and Freeport should tell us whether the industrial recovery is showing up in orders and physical demand. GE Vernova should tell us whether the power buildout is still moving from announcement to execution, while oil will probably continue reacting to missiles and ceasefire proposals in roughly equal measure.
For now, I’m not chasing the sudden discovery that semiconductor stocks became crowded, and I’m not changing the energy view around every diplomatic headline. The better approach is to let the earnings and physical-market data resolve the contradictions for us.
The market has several stories it would like us to believe. This week, we should find out which ones still have numbers behind them.
Luke Perry
Portfolio Manager, Whalen Financial