Markets have just delivered one of the strongest six-week rallies in years, with the S&P holding well above 7,400 even on days when the board is red across the board. The natural question is whether this run has further to go. The honest answer is that while momentum has driven much of this year's gains, the move is not built on momentum alone. There is genuine fundamental support beneath it: double-digit earnings growth across the broader sectors, paired with some of the highest beat-and-raise rates seen in years.
That said, valuation deserves respect. The market stretched to roughly 24 times earnings earlier in the year, gave some of that back in March, and now sits closer to a 22 times multiple. From a momentum standpoint, blow-off risk is real. This is precisely why discipline and diversification matter more than chasing the tape. Opportunity in momentum names is real too — the goal is to participate without becoming dangerously concentrated.
A Diversified Approach Across Themes
The most durable way to engage this environment is to spread exposure across several reinforcing themes rather than betting on a single trade.
One of the earliest and most compelling is electrification — the build-out of physical infrastructure required to support AI. The power, grid, and electrical-equipment names tied to this theme benefit directly from surging AI energy demand. There are strong ETFs covering the space, and individual leaders such as Eaton sit squarely in that current.
A second pillar is natural resources, which serve a dual purpose: a geopolitical hedge on one hand, and a play on the relentless AI energy demand on the other. A third is semiconductors and memory chips, the dominant story of recent months, where measured exposure is warranted. A fourth is the re-rating of large-cap software, where some of the biggest technology names are being repriced from growth to value in real time. Finally, international companies stand to benefit as rising fiscal spending abroad provides a tailwind that is largely independent of the U.S. cycle.
The point of holding all of these together is resilience: each leans on a different driver, so the portfolio is not hostage to any one narrative.
Software: The "SaaS-pocalypse" Is Overblown
Few corners of the market have been as emotionally traded as enterprise software. The fear — call it the "SaaS-pocalypse" — is that AI will hollow out software business models almost overnight. The extended software ETF, IGV, sold off sharply on exactly this anxiety earlier in the year, which made it an attractive entry point.
The bearish thesis underestimates how software is actually consumed inside organizations. Many of these companies are industry leaders in enterprise software, with deeply ingrained workflows and inherently high switching costs. The market has been pricing them as though they will never grow again and their models will evaporate immediately. In practice, that is implausible. No serious leadership team is going to entrust mission-critical enterprise infrastructure to something a team member casually "vibe coded" together and hope it holds up in production. The trust, integration, and continuity requirements are simply too high. That is why software is likely to re-rate higher from here rather than collapse.
Semiconductors: An Arms Race With Room to Run
On the other side sits the semiconductor complex, which has shifted from a period of consolidation into what looks like unstoppable, earnings-driven growth. Whether the trade is now overcrowded or still has room is genuinely hard to call. But the structural logic is straightforward: follow the money. The enormous CapEx flowing from the hyperscalers is funneled almost entirely back into chips — Nvidia's GPUs, Taiwan Semiconductor's foundry capacity, and Broadcom's custom silicon. There is little leakage; the spending lands directly on the semiconductor leaders paving the front edge of AI infrastructure demand.
The open question remains return on investment for the hyperscalers themselves. But at this point it is effectively an arms race, and in an arms race the suppliers of the weapons keep selling. On balance, there is still room for many of these names to run.
Nvidia: The Stalwart Into Earnings
Nvidia deserves its own consideration as the industry's bellwether and darling. As Nvidia goes, so goes the broader AI infrastructure trade and, to a large extent, the market itself. Heading into its next report, the setup is different from before: after a long consolidated trading range, the stock has broken out, helped by reports that it will be permitted to sell H200 chips to select customers in China.
The recurring pattern, however, is worth noting — blockbuster report after blockbuster report met by a tepid reaction. There was a stretch where even strong Nvidia numbers could not push the S&P back above 7,000. Some of this is simply the law of large numbers. When a company is worth more than five trillion dollars, the market cannot reasonably expect its market cap to jump ten percent overnight, no matter how good the quarter. Conviction in the name remains high: GPU demand is intense, and the company has extended its reach across the broader AI stack — fiber optics, optical and interconnect technology, and neo-cloud investments — embedding itself well beyond raw chips.
When Reaction Matters More Than the Numbers
This points to perhaps the most interesting dynamic of the current cycle. Nvidia is the clearest example, but not the only one: companies posting beats and raises have been selling off ten percent on the news. We may have entered a phase where investor reaction matters as much as — or even more than — the actual results.
That appears to be the case. Investor psychology has become paramount to how markets move, with leading names swinging violently day to day in ways the fundamentals alone do not explain. Understanding this environment increasingly requires thinking about market psychology as seriously as earnings models. The numbers still matter — but how investors choose to feel about those numbers has become its own decisive variable.