AI Dividend: Earnings Revisions Drive Tech Strength Beyond Momentum

The current tech surge, while impressive, reveals a market increasingly bifurcated by AI adoption and a surprising resilience in semiconductor demand, driven by long-term capital expenditure cycles that extend beyond immediate market sentiment. This conversation with Pete Callahan, US Technology, Media and Telecommunications sector specialist at Goldman Sachs, highlights that the rally's narrow breadth and high velocity mask underlying strength in earnings revisions, particularly in AI-driven sectors. Investors who understand the extended visibility into 2027 for AI infrastructure companies and can navigate the dispersion within software--where AI deployment is the key differentiator--will find opportunities. This analysis is crucial for investors seeking to move beyond headline numbers and identify durable competitive advantages in a rapidly evolving tech landscape, offering a distinct edge over those focused solely on short-term market movements.

The AI Dividend: Why Earnings Revisions Trump Momentum

The recent tech rally, characterized by its rapid ascent and narrow leadership, might feel like a familiar, perhaps even precarious, ascent. However, digging beneath the surface, as Pete Callahan explains, reveals a more nuanced picture driven by fundamental earnings growth and revisions, particularly within the AI ecosystem. While the Nasdaq has surged nearly 20% off its March lows, the breadth of this rally is surprisingly limited, with only about half of its constituent stocks participating. This isn't a broad-based market euphoria; it's a concentrated push, heavily anchored by semiconductors, which have seen an astonishing 80% rise this year--their best performance since 1999.

This isn't just about price appreciation; it's about the underlying economics. Callahan points out that AI stocks within the S&P 500, up approximately 30% year-to-date, have seen their earnings rise in lockstep. This suggests that multiples are being kept in check by robust earnings revisions, a critical distinction from speculative bubbles. The narrative here is one of companies delivering tangible results that justify their valuations, rather than simply riding a wave of market exuberance.

The implications for investors are significant. Instead of chasing broad market trends, the focus must shift to identifying companies that are not only participating in the AI revolution but are also demonstrably benefiting from it in terms of earnings. This requires a deeper dive into earnings revisions and an understanding of which sectors are experiencing genuine, sustained demand.

"What do I mean by that? If I look back over the last three months, the Nasdaq is up 20%. But when you look under the hood, only about half of the stocks in the Nasdaq are even up at all over that stretch. So you have narrow breadth, again, led by semiconductors, which are up nearly 80% this year, their best year since 1999."

The semiconductor sector, in particular, exemplifies this dynamic. Its exceptional performance is not merely a function of investor sentiment but is underpinned by significant capital expenditure. Callahan highlights that CapEx estimates for calendar 2027 have been revised upward by 20%, projecting over $900 billion in spending. This long-term investment cycle suggests that the demand for AI infrastructure is not a fleeting trend but a sustained build-out. Companies that are integral to this supply chain, even if they seem like obvious beneficiaries, are positioned for continued growth as long as these CapEx trends persist.

Software's AI Crucible: Dispersion and Differentiation

Within the software sector, the story is more complex, marked by a welcome dispersion after a period of undifferentiated pressure. While the broader market might have worried about AI disintermediating software companies, the reality is proving more nuanced. Callahan notes that the key differentiator for software companies now is their ability to integrate and deploy AI effectively, which in turn drives faster revenue growth and justifies higher market valuations.

"It was interesting because everybody wanted to sell software because they thought they would be disintermediated by AI, which is software."

This presents a critical insight: the companies that are actively leveraging AI to enhance their existing offerings, rather than being replaced by it, are the ones capturing investor attention. The debate has shifted from "will AI kill software?" to "which software companies are best positioned to harness AI?" This requires investors to look beyond the general category of "software" and assess individual companies based on their AI strategy, product innovation, and demonstrated impact on revenue.

The ongoing tension between subscription and consumption models, and the incumbent versus startup dynamic, are all playing out against the backdrop of AI integration. Companies that can clearly articulate how AI is enhancing their value proposition and driving customer adoption are likely to outperform. This focus on AI deployment as a driver of differentiation is a crucial element for understanding the software market's future trajectory, moving beyond the correlation-one pressures of the past.

The Long Game: Semiconductor Resilience and Extended Visibility

The exceptional performance of semiconductors, while reminiscent of the dot-com bubble's peak, is grounded in a different reality. Callahan acknowledges the short-term concerns about momentum and "too far, too fast" dynamics when a sector rises 80% in five months. However, he pivots to the medium-term driver: earnings revisions. As long as these revisions continue, providing a fundamental anchor for valuations, investors will remain comfortable adding to positions on pullbacks.

The critical takeaway here is the extended visibility into calendar 2027 for semiconductor and AI infrastructure companies. This suggests that the current demand is not a short-term spike but part of a multi-year build-out. Companies are securing orders that extend far into the future, driven by persistent shortages and the ongoing need for compute power and AI infrastructure.

This extended visibility creates a significant competitive advantage for those who can secure capacity and deliver reliably. It also means that traditional valuation metrics might need to be viewed through a longer-term lens. The "short-term pain for long-term gain" ethos is evident here, as companies invest heavily in capacity and R&D to meet future demand, even if it means managing tight supply and potentially higher input costs in the present.

Where Conventional Wisdom Fails

Conventional wisdom might suggest caution given the rapid ascent of semiconductors. However, Callahan's analysis implies that for those companies with genuine demand extending years out, and for investors willing to look past short-term fluctuations, the opportunity remains. The system's response--the sustained CapEx and extended order books--indicates that the demand is deeply embedded. The failure of conventional wisdom lies in applying short-term momentum analysis to a market driven by long-term infrastructure build-out.

The US Internet Sector's AI Pivot

While semiconductors and AI-driven software have captured headlines, the US Internet sector has lagged, a surprising development given its consumer-facing nature. Callahan attributes this to ongoing debates surrounding funding sources, investment cycles, and consumer health, alongside the evolving role of AI in the consumer space.

However, he signals a potential shift. Recent product innovation tied to AI within US Internet companies, coupled with moderating consumer spending (partially due to resetting oil prices), suggests a cleaner positioning for the sector. This indicates that AI is not just an enterprise play but is also beginning to influence consumer-facing internet businesses.

This presents an opportunity for investors who recognize that AI's impact is broad, extending beyond the obvious infrastructure providers. Companies that can effectively integrate AI into their consumer products and services, and whose business models are resilient to shifts in consumer spending, could see a resurgence. The "discomfort now creates advantage later" principle applies here, as companies investing in AI integration may face upfront costs but stand to gain significant market share and customer loyalty in the long run.

Navigating the Macro and Tech Landscape

Looking ahead to June, Callahan emphasizes the importance of macro data, specifically Non-Farm Payrolls (NFP) and Consumer Price Index (CPI) reports, for understanding inflation and the broader economic environment. These will provide crucial context for interest rate expectations, which historically have influenced tech valuations.

However, he notes that rates have not been a primary concern for tech recently. This is largely due to the significant cost inflation inputs already weathered by the AI supply chain. The market seems to be pricing in these input costs and focusing on the delivery of AI infrastructure. This suggests a degree of resilience in tech valuations, even in a higher-rate environment, as long as earnings growth continues to outpace cost increases.

The upcoming user conferences in software and semiconductors will also be key for gauging the sentiment and trajectory of generative AI development into the summer. These events will offer insights into product roadmaps, customer adoption trends, and the overall health of the AI ecosystem.

Key Action Items

• Immediate Action: Analyze the breadth of the current tech rally. Identify which companies are participating beyond the obvious AI leaders and assess their underlying earnings growth.
• Immediate Action: Within software, focus on companies demonstrating clear AI integration and its impact on revenue growth. Prioritize those with strong subscription or consumption models that are being enhanced by AI.
• Immediate Action: Track CapEx revisions for calendar 2027 in the semiconductor and AI infrastructure sectors. This provides a longer-term indicator of sustained demand.
• Over the next quarter: Evaluate US Internet companies for AI-driven product innovation and their resilience to consumer spending fluctuations.
• Over the next 12-18 months: Monitor user conferences and industry events for insights into generative AI's continued evolution and adoption across different tech segments.
• Longer-term investment: Consider companies with extended visibility into future demand (e.g., into 2027 for AI infrastructure), as this indicates a more durable competitive advantage.
• Strategic consideration: Be prepared to invest in sectors or companies that require significant upfront investment or present short-term challenges but offer substantial long-term payoffs due to AI integration and infrastructure build-out.