The Certainty Premium: What’s Driving the Tech Sector Divergence? | Jeff Keller

Jeff Keller analyzes the tech sector's current landscape, emphasizing that long-term uncertainty is creating a premium on short-term certainty. He observes a disconnect between financial analysis and the optimistic outlook in Silicon Valley, with the market prioritizing growth potential over individual market share. Keller notes a 20% increase in software and services since Q3, while technology hardware remains stagnant and semiconductors have dropped by 7.5%.

He warns against extrapolating short-term trends, highlighting that semiconductor companies are trading at varying multiples despite similar fundamentals. Keller predicts that lower multiple names in the AI sector may outperform more speculative firms that are pricing in overly optimistic futures. He expresses skepticism about the economics behind recent tech announcements, indicating that the market is becoming more cautious and focused on sustainable profits.

Keller discusses the impact of rising long-end bond yields and financial tightening on companies' capital-raising abilities. Despite these challenges, major tech firms have raised around $500 billion year to date, with many still achieving double-digit growth. He notes that the hyperscaler model that once drove performance for major tech companies is evolving due to increased competition, raising questions about the sustainability of growth in the sector.

He points out that while the return on investment for hyperscalers remains positive, market perceptions can shift quickly. Keller criticizes Meta for its slow cost-cutting response amid stock price pressures and suggests that the company's messaging on compute capacity serves more as a signal to investors than a reflection of its business impact. He believes Meta's challenges do not necessarily indicate broader issues in the AI sector.

Keller predicts continued data center construction beyond the midterms, citing independent growth stories in e-commerce and recent IPOs as promising opportunities. He describes his investment fund as long biased, with over 50% net exposure, and views a long bias model with extended duration as appealing in the current environment. He reflects on the overly optimistic sentiment among tech investors in 2021, which led to significant losses, and asserts that current market conditions do not warrant a fully net short position on tech.

He warns that the AI trade may eventually mirror the conditions seen in late 2021, while acknowledging that AI will not eliminate market emotions like fear and greed. Keller observes that the speed of market theme recognition is increasing, making awareness of factor exposures more critical. He notes that AI funds may have contributed to negative momentum in stocks, resulting in one of the worst five-day stretches for momentum in two decades.

Keller identifies opportunities when sectors are deemed uninvestable, citing margin calls among the Korean population as evidence of substantial deleveraging. He emphasizes the importance of optionality in portfolio management and predicts that shorter time horizons and increased uncertainty will become more common. He acknowledges pronounced volatility in certain sectors, even as headline indices appear stable, and expresses uncertainty about whether market movements are driven by fundamentals or technical factors.

He anticipates that the upcoming Anthropic S1 will provide metrics to clarify the market situation, with key factors to watch including annual recurring revenue and capital expenditures. Keller suggests that stocks may fluctuate following recent retail enthusiasm and acknowledges the potential for technology breakthroughs to reignite investor excitement, though the timing remains uncertain. He mentions the possibility of certain areas within the AI trade experiencing a "SpaceX effect," while distinguishing current equity issuance from past speculative bubbles. He concludes that AI stocks could rise leading up to the Anthropic IPO before potentially declining, noting that euphoria in the semiconductor space has diminished.