A Tale of Two Software Companies
The most instructive market story right now is not whether artificial intelligence will reshape enterprise software — that much is taken for granted — but rather which incumbents emerge stronger and which get hollowed out. Two large software franchises illustrate the divergence cleanly, and the fact that the same analytical house reached opposite conclusions about them on the same day underscores how granular the AI thesis has become.
ServiceNow is the optimistic case. After a punishing stretch in which the stock fell roughly 52% over the trailing year, a fresh reinstatement of coverage with a buy rating and a $130 price target — implying more than 30% upside — sent shares up over 4% in a single session. The argument is structural rather than tactical. Far from being disrupted by AI, ServiceNow is seen as positioned to benefit precisely because it sits so deeply embedded at the center of enterprise workflows. As autonomous AI agents proliferate inside corporations, something has to govern them: approvals, permissions, audit trails. ServiceNow is cast as that connective and control layer, the place where agentic activity is orchestrated and supervised. The accompanying financial picture supports the conviction — revenue growth projected in the 18% to 22% range through 2028, free cash flow margins in the mid-30% range, and a valuation that now looks attractive relative to peers after the share-price decline. In short, AI is framed as an opportunity, not a threat.
Salesforce is the cautionary counterpoint. Coverage there was reinstated with an underperform rating and a $160 price target. The concern is not a temporary air pocket but a structural reset — a cyclical slowdown layered on top of AI fundamentally reshaping how businesses operate. The nuance matters: Salesforce's Agentforce product is growing, but it is growing largely within an already deeply entrenched base of existing customers. There is limited room to expand beyond that installed base, and the longer-term worry is more corrosive. Because Salesforce monetizes on a per-seat (software-as-a-service) model, the very efficiency that AI delivers — fewer humans needed to do the same work — could translate into a reduction in paid users over time. The same force that makes ServiceNow a toll booth on AI activity makes Salesforce's seat-based pricing a liability. The market reaction was telling: shares were little changed, suggesting investors are still digesting whether the bear case is right.
The broader lesson is that "AI disruption" is not a uniform tide that lifts or sinks software wholesale. The decisive variable is where in the stack a company sits and how it charges. Businesses that govern, route, and audit AI activity capture value as that activity grows; businesses whose revenue is mechanically tied to human headcount face a deflationary headwind when AI reduces the need for those humans.
The Chip Equipment Cycle and the DRAM Question
A parallel sorting exercise is underway in semiconductor capital equipment, where the overall thesis is bullish but the security selection is sharpening. The chip-buildout story tied to the AI boom is intact, and industry spending forecasts have been raised. The interesting move is relative rather than directional: Lam Research was upgraded to overweight while Applied Materials was downgraded to equal weight — a "who's in, who's out" call within a sector that is broadly favored.
Applied Materials is not being abandoned. It is still regarded as a solid business with likely gains in 2026. But its growth is expected to track the broader market, whereas Lam Research is identified as the company better positioned for the next leg of growth. The crux is the memory market, specifically the dynamic of DRAM versus NAND, where demand continues to outpace supply. This imbalance is understood not as a one-quarter phenomenon but as a multi-year spending cycle — a view shared across more than one analytical house, not an outlier call. The key swing factor is the timing and strength of a NAND recovery, and Lam Research is judged to carry more leverage to NAND equipment than its rivals. That greater leverage to the recovering part of the memory complex is what tilts the outperformance case toward Lam.
This is the same intellectual move as the software story in different clothing: a secular tailwind everyone agrees on, with returns concentrated in the names most exposed to the specific sub-segment doing the heavy lifting.
Berkshire's Filing: A Quiet Return and a Generational Handoff
The latest regulatory disclosure of portfolio holdings from Berkshire Hathaway offered its usual parlor game of "who's in and who's out," and several moves stand out.
The headline is Delta. Berkshire revealed a new stake of roughly $2.6 billion in the airline, and the shares moved higher on the disclosure. The significance lies in the history. In the early days of the pandemic, Warren Buffett had sold Berkshire's entire airline position — billions of dollars spanning United, American, Southwest, and Delta — explaining at the time that the pandemic had altered consumer behavior and what travel would look like. Exiting the sector entirely and then returning is therefore a notable reversal. While Delta now ranks only as Berkshire's 14th-largest holding — hardly the top of the book — the symbolism of re-entering a space the firm had publicly walked away from is what gives the move its weight. It also dovetails with a popular analyst framing of the so-called K-shaped consumer: Delta, with its premium pricing and premium customer base, is seen as a beneficiary of an economy where higher-income households continue to spend even as lower-income ones retrench.
Other changes round out the picture. Berkshire more than tripled its position in Alphabet, vaulting it to the conglomerate's seventh-largest holding — a meaningful endorsement of the technology platform. A new, smaller stake in Macy's was disclosed. On the exit side, Berkshire sold out of UnitedHealth, a position it had bought into during a sharp selloff last year for roughly $5 million worth of stock; that holding is now gone.
Underneath the individual trades sits a larger transition. Investors have long aspired to "invest like Buffett," but the operative phrase is shifting toward investing like Greg Abel — or simply like Berkshire as an institution — even as Buffett remains heavily involved. The lesson embedded in these filings is less about copying any single position and more about the temperament on display: a willingness to reverse a high-profile decision when conditions change, to concentrate into conviction (Alphabet), to take small exploratory positions (Macy's), and to exit cleanly without sentiment (UnitedHealth). The discipline, not the ticker list, is the thing worth imitating.
The Common Thread
Across software, semiconductor equipment, and one of the world's most-watched portfolios, the same principle recurs: broad theses are cheap, and differentiation is everything. AI will reshape enterprise software — but it rewards the governing layer and punishes seat-based pricing. The chip buildout is real — but returns flow to the names levered to the recovering sub-segment. And legendary investing is not a static set of holdings — it is the willingness to change one's mind when the facts change. In each case, the money is made not by being right about the trend, but by being precise about who within it actually captures the value.