The artificial intelligence build-out has been narrated mostly through the lens of the headline chip designers, but the more durable story is unfolding one layer down — among the companies that make the machines that make the chips, and the companies that store the data those chips generate. In recent trading, this quieter cohort has held up remarkably well even as the broader semiconductor complex has churned through bouts of volatility. That divergence is worth understanding, because it says something important about where the real, contracted demand in this cycle actually sits.
The Equipment Makers Sit Above the Volatility
The companies that supply wafer-fabrication equipment — the tools used to manufacture semiconductors — have proven notably less sensitive to the swings that have rattled chip stocks over the past several weeks. Names like Lam Research, KLA-Tencor, and Applied Materials have moved steadily higher rather than lurching with sentiment. The reason is structural: their fortunes are tied to the capital-expenditure cycle of their customers, and that cycle is strengthening across the board.
Analyst conviction reflects this. Barclays has continued to push its price objectives higher across the group, citing improving demand and a stronger CapEx environment. The firm lifted its target on KLA-Tencor substantially — to $2,250 from $1,700 — while raising Lam Research, a large memory-equipment maker, to $355 from $275, and Applied Materials to $590 from $500. These are not modest revisions. They reflect a view that sales will keep ramping not just into next year but through 2027 and 2028, underpinned by demand from AI, data centers, memory, and a CPU market that is itself running hot right now.
The logic here is the classic "picks and shovels" insight. When demand for computation surges, the spending flows first to those building capacity. Equipment makers capture that spending early in the chain and aggregate it across every customer expanding a fab, which insulates them somewhat from the fortunes of any single chip designer. That diversification is precisely why they have been less subject to the recent semiconductor volatility.
Storage Rides the Same Wave
The same demand dynamics are visible in storage. Seagate has been a remarkable performer, with shares up more than 500% over the past year — a move driven by its position in the memory stack across both NAND flash and DRAM. The thesis is straightforward: major hyperscalers building out AI infrastructure need both memory and storage in enormous quantities, and that need is not abating.
JP Morgan reinforced the view by raising its price target on Seagate to $920 from $775 while maintaining an overweight rating. The case rests on improving margins, robust order books extending into 2028 on consensus estimates, and continued accretive contributions to sales and earnings. Notably, this move followed other constructive analyst revisions on the name, suggesting a broadening consensus rather than a single optimistic outlier. Storage, in other words, is being repriced as a structural beneficiary of the hyperscaler capital cycle rather than a commoditized afterthought.
Where the Cycle Meets Its Limits: The Cautionary Case
Not every company touched by the AI build-out enjoys the same clean narrative, and the contrast is instructive. Super Micro illustrates how company-specific risk can override an otherwise favorable demand backdrop. After a sharp sell-off, the stock managed only the faintest recovery — a fraction of a percent — and that flicker faded almost as quickly as it appeared. It will take far more than a marginal bounce to repair the damage from a decline of that magnitude.
The immediate trigger was equity dilution. The company raised roughly $7 billion through a combination of common stock and convertible preferreds — a sum equal to about a third of its market capitalization, which sat near $22 billion. The proceeds are meant to help fund a server build-out on the order of $39 billion, but the market punished the dilution heavily, especially as it landed in an already weaker tape with server peers also under pressure.
Underneath the dilution lies a deeper set of concerns. The company carries governance baggage dating back several years, including accounting irregularities, and faces a Department of Justice investigation tied to allegedly bypassing export constraints on GPU technology routed through its servers. Wolf Research initiated coverage at a peer-perform rating, reiterating these risks alongside others — notably heavy customer concentration, the dilution itself, and the overhang of the DOJ indictment. Even established competitors in the server space, including Hewlett Packard Enterprise and Dell, have been pulling back, compounding the weakness across the segment.
The Lesson Beneath the Tickers
Taken together, these three cases sketch a clear hierarchy of exposure within the AI infrastructure trade. The equipment makers occupy the most defensible ground, capturing diversified CapEx spending early and proving resilient to sentiment shocks. Storage suppliers ride a similar structural tailwind, repriced upward as hyperscaler demand proves durable into the latter half of the decade. Server builders, by contrast, sit closer to the customer and are more exposed — both to the capital intensity of building out capacity, which can force dilutive financing, and to idiosyncratic governance and regulatory risk.
The broader takeaway is that "AI demand" is not a single, uniform tide that lifts every related stock equally. The strongest positions belong to those whose revenue is contracted years out, whose customer base is diversified, and whose balance sheets do not require punishing the existing shareholder base to fund growth. In a cycle this large, the durable winners may not be the most visible names, but the ones quietly selling the tools and the storage that make the whole edifice possible.