Macro Pressure Overwhelming Strong Micro Fundamentals
The current market is defined by a rapid-fire rotation happening beneath the surface — between hardware, software, and the Mag 7 names. This churn makes sense once you recognize what is driving it: a wave of macro pressure that is temporarily overweighing genuinely strong fundamental earnings on the micro side. The result is a burst of short-term volatility that has little to do with the underlying health of the businesses involved.
Several forces are converging at once to create this noise. Elevated oil prices stemming from the situation in Iran carry broad implications, but they matter in a specific and underappreciated way for the "jewels" of the AI supply chain — South Korea and Taiwan. Both regions are largely energy dependent, so higher energy costs put pressure on their currencies and stoke inflation, adding another layer of volatility to the very geographies that feed critical components into the US AI build-out.
Rather than viewing this volatility as a threat, it is better understood as a reason for investors to pause, step back, and study the full AI picture — to decide where within the AI landscape they actually want to be positioned. The key screening question becomes: which companies can take the exposure they've gained from AI and convert it into actual, durable earnings power?
Memory as the Proof Point of Durable Value
The clearest place this durable value is solidifying is in the high-bandwidth memory space, specifically the DRAM segment. This crystallized last week with Micron's earnings — a genuinely strong quarter with strong results and strong guidance. To understand why memory accrues so much value, follow the money: of the roughly one trillion dollars of CapEx flowing into US data centers, nearly 40% is going into memory. That concentration is precisely why memory is not merely AI-exposed but is actually capturing a large share of the durable value being created. Memory is therefore the "proof point" to look for — the segment where AI spending translates most reliably into lasting earnings.
Memory is a well-known story that most investors are already cognizant of. A more interesting, less crowded angle lies in the proliferation of data itself. Data has to be stored, but it also has to be moved — which points toward networking plays and interconnects. Two companies in particular stand out:
- Marvell, on the networking fabric side.
- Semtech, which has strong positioning centered on signal integrity.
The logic behind Semtech is that moving data is only half the challenge; maintaining the highest degree of signal integrity while moving it is equally important, because that is what preserves low latency, low power draw, and the fastest possible speeds.
Check Writers Versus Check Takers
The migration of capital away from the hyperscalers and into the bottlenecks — memory, AI infrastructure, and networking — has been characterized as a shift from the "check writers" to the "check takers." A telling episode unfolded in South Korea, where the check takers, SK Hynix and Samsung, effectively became check writers themselves. Capital rotated out of those very well-performing names and into industrials and renewable energy, on the expectation that the fabs they plan to build out in Korea will be powered by renewable energy.
A parallel question is whether the same dynamic will play out in the US: the check beneficiaries who are taking advantage of all that CapEx are themselves now spending heavily, and industrials have been performing strongly month to date, suggesting a rotation into that sector could follow.
Within the US AI trade, part of what has occurred is a rotation out — investors taking profits and moving into the more defensive positions available in the current market. But the more significant story is the rotation within AI itself. Interconnect, optics, and cooling companies performed well through the first two quarters of the year, but they have begun lagging the memory companies. That divergence reflects investors stepping back to identify who actually owns the long-term durable value, which has channeled more investment toward the memory names and toward select semiconductor companies.
How to Invest Around a Flip-Flopping AI Trade
The rotation under the surface has become dizzying — some mornings all the memory names are down while everything else is up; some days hardware leads, other days software leads; sometimes the Mag 7 rises while the broader market falls; and on other days everything rallies in unison. This raises a natural question: with such a short-term attention span and constant seesawing beneath the surface, is it better to capture the whole AI theme through a single index rather than be selective in a particular category that could reverse the next day?
The answer is to focus, again, on who is actually capturing the value. Over the past year, rising tides lifted all boats — nearly every AI-exposed name enjoyed a significant run. That indiscriminate phase is now changing. Going forward, investors will want more exposure to the names that can provide durable franchise value based on where they sit within the infrastructure stack. The short-term volatility should be viewed more as opportunity than as risk: the macro noise is not breaking the long-term AI build-out; it is merely creating a more difficult short-term trading setup. For names in which an investor has high long-term conviction, these short-term dislocations can be a genuinely good buying opportunity.
Where the Macro Actually Weighs on the Micro
There is an apparent tension worth confronting. Year-end targets have been built on relentless earnings strength — the market has experienced a positive earnings shock, largely driven by tech. There is debate about the narrowness of that earnings profile, though some argue, encouragingly, that it is broadening out. And this is a market that rallied straight through an entire war and still hit record highs. So where, exactly, is the macro weighing on the micro — and could the AI-driven inflation flagged by Apple become a real problem?
The pressure comes from an ever-growing list of factors piling onto volatility:
- Fed uncertainty — investors are trying to digest the Fed decision and grow comfortable with a potentially more hawkish stance, all while having less visibility into the path ahead.
- The Iran conflict — and how it feeds through to energy prices and inflation, given that the critical, geographically concentrated supply chains feeding the US AI build-out sit in energy-dependent regions.
- Memory inflation — which is beginning to trickle through and itself feeds back into the Fed's calculus, raising the question of whether it will drive specific policy adjustments.
With all of these converging at once, it was easy for the macro noise to quickly overwhelm the otherwise strong Micron earnings.
What the Next 30 Days Hold
Looking to the short term — the next 30 days, toward the end of July — a major earnings season is coming that will restore visibility. Key semiconductor players such as ASML and TSMC are set to report, offering much clearer insight into capacity lead times and whether these players are leaning into demand. The hyperscalers also report toward the end of the month. If the other companies capturing durable value echo what Micron delivered, this earnings season represents a strong opportunity for the micro to push back against the macro — for solid fundamentals to reassert themselves over the short-term noise.
