UBS published a client note on Monday that uses lessons from earlier technology cycles to advise investors amid the current AI-driven rally. The bank distilled three core principles intended to help market participants position portfolios as technology valuations climb and spending on AI infrastructure accelerates.
Markets have already reflected heavy investor interest in technology. The Nasdaq has risen by more than 15% over the past month, and the Philadelphia Semiconductor index has increased by nearly 70% since the start of April, according to the data cited in the note.
On spending, UBS reiterated outside estimates that AI-related capital expenditure is expected to reach $820 billion in 2026 and approach $990 billion in 2027. The bank noted that more than 85% of that projected spending will be driven by the big four technology companies, underscoring concentration of demand at the top of the market.
UBS's first principle is that incumbents do not automatically secure leadership in the next technological era. The note points to historical examples such as IBM's inability to dominate the personal computer era as an illustration of how established firms can be slow to adapt when paradigms shift.
The second principle emphasizes that durable economic value tends to migrate toward platforms rather than the builders of enabling infrastructure. UBS framed this as a dynamic that favors companies capable of creating sticky user ecosystems and monetizable applications, rather than firms primarily supplying hardware or other enabling components.
The third lesson addresses the lifecycle of capital expenditure booms. UBS cautioned that when capacity eventually catches up with demand, segments focused on hardware typically face the first and most acute pressure. The bank cited Nortel's bankruptcy and the long path Cisco took to return to its dotcom-era highs - a roughly 25-year span - as examples that illustrate the risks hardware suppliers can encounter after a period of rapid investment.
On portfolio construction, UBS said it believes owning both the intelligence layer and the applications layer will be important going forward. The bank recommended maintaining diversification across AI layers rather than concentrating holdings solely in a narrow group of mega-cap names.
These recommendations reflect an emphasis on resilience in business models - sticky revenue streams from platforms, monetizable applications and the ability to adapt underwriting of investment in capacity - while highlighting the vulnerability of hardware-focused segments when capex cycles normalize.
Bottom line: UBS's note uses historical parallels to argue for diversified exposure across AI ecosystem layers, warns investors that dominant firms today can be displaced, and flags hardware suppliers as particularly exposed when the current investment surge subsides.